An AI-powered Army of Darkness It's game night, your friends are perched around the games table, waiting to see what Dungeons & Dragons (D&D) character they'll become and quest they'll embark on. Tonight, you're Dungeon Master (storyteller and guide), crafter of thrilling encounters to challenge and enthrall your players. Your trusty D&D Monster Manual contains thousands of creatures. Finding the perfect monster for each situation among the myriad options can be overwhelming. The ideal foe needs to match the setting, difficulty, and narrative of the moment. What if we could create a tool that instantly finds the monster most suited to each scenario? A tool that considers multiple factors simultaneously, ensuring each encounter is as immersive and exciting as possible? Let's embark on a quest of our own: build the ultimate monster-finding system, using the power of multi-attribute vector search! Creating creatures with vector search, why do it? Vector search represents a revolution in information retrieval. Vector embedding - by taking account of context and semantic meaning - empowers vector search to return more relevant and accurate results, handle not just structured but also unstructured data and multiple languages, and scale. But to generate high quality responses in real-world applications, we often need to assign different weights to specific attributes of our data objects. There are two common approaches to multi-attribute vector search. Both start by separately embedding each attribute of a data object. The main difference between these two approaches is in how our embeddings are stored and searched. the naive approach - store each attribute vector in separate vector stores (one per attribute), perform a separate search for each attribute, combine search results, and post-process (e.g., weight) as required. the Superlinked approach - concatenate and store all attribute vectors in the same vector store (using Superlinked's built-in funtionality), which allows us to search just once, with attendant efficiency gains. Superlinked's spaces also let us weight each attribute at query time to surface more relevant results, with no post-processing. Below, we'll use these two approaches to implement a multi-attribute vector search tool - a Dungeons and Dragons monster finder! Our simple implementations, especially the second, will illustrate how to create more powerful and flexible search systems, ones that can handle complex, multi-faceted queries with ease, whatever your use case. If you're new to vector similarity search, don't worry! We've got you covered - check out our building blocks articles. Okay, let's go monster hunting! Dataset First, we'll generate a small synthetic dataset of monsters, by prompting a Large Language Model (LLM): Generate two JSON lists: 'monsters' and 'queries'. 1. 'monsters' list: Create 20 unique monsters with the following properties: - name: A distinctive name - look: Brief description of appearance (2-3 sentences) - habitat: Where the monster lives (2-3 sentences) - behavior: How the monster acts (2-3 sentences) Ensure some monsters share similar features while remaining distinct. 2. 'queries' list: Create 5 queries to search for monsters: - Each query should be in the format: {look: "...", habitat: "...", behavior: "..."} - Use simple, brief descriptions (1-3 words per field) - Make queries somewhat general to match multiple monsters Output format: { "monsters": [ {"name": "...", "look": "...", "habitat": "...", "behavior": "..."}, ... ], "queries": [ {"look": "...", "habitat": "...", "behavior": "..."}, ... ] } Let's take a look at a sample of the dataset our LLM generated. Note: LLM generation is non-deterministic, so your results may differ. Here are our first five monsters: # name look habitat behavior 0 Luminoth Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey 1 Aqua Wraith Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents 2 Stoneheart Golem Massive humanoid composed of interlocking rock formations Rocky mountains and ancient ruins Hibernates for centuries, awakens to protect its territory 3 Whispering Shade Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths 4 Zephyr Dancer Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates ...and our generated queries: Look Habitat Behavior 0 Glowing Dark places Light manipulation 1 Elemental Extreme environments Environmental control 2 Shapeshifting Varied landscapes Illusion creation 3 Crystalline Mineral-rich areas Energy absorption 4 Ethereal Atmospheric Mind influence See original dataset and query examples here. Retrieval Let's set up parameters we'll use in both of our approaches - naive and Superlinked - below. We generate our vector embeddings with: sentence-transformers/all-mpnet-base-v2. For simplicity's sake, we'll limit our output to the top 3 matches. (For complete code, including necessary imports and helper functions, see the notebook.) LIMIT = 3 MODEL_NAME = "sentence-transformers/all-mpnet-base-v2" Now, let's get our multi-attribute monster search under way! First, we'll try the naive approach. Naive approach In our naive approach, we embed attributes independently and store them in different indices. At query time, we run multiple kNN-searches on all the indices, and then combine all our partial results into one. We start by defining a class NaiveRetriever to perform similarity-based search on our dataset, using our all-mpnet-base-v2-generated embeddings. class NaiveRetriever: def __init__(self, data: pd.DataFrame): self.model = SentenceTransformer(MODEL_NAME) self.data = data.copy() self.ids = self.data.index.to_list() self.knns = {} for key in self.data: embeddings = self.model.encode(self.data[key].values) knn = NearestNeighbors(metric="cosine").fit(embeddings) self.knns[key] = knn def search_key(self, key: str, value: str, limit: int = LIMIT) -> pd.DataFrame: embedding = self.model.encode(value) knn = self.knns[key] distances, indices = knn.kneighbors( [embedding], n_neighbors=limit, return_distance=True ) ids = [self.ids[i] for i in indices[0]] similarities = (1 - distances).flatten() # by definition: # cosine distance = 1 - cosine similarity result = pd.DataFrame( {"id": ids, f"score_{key}": similarities, key: self.data[key][ids]} ) result.set_index("id", inplace=True) return result def search(self, query: dict, limit: int = LIMIT) -> pd.DataFrame: results = [] for key, value in query.items(): if key not in self.knns: continue result_key = self.search_key(key, value, limit=limit) result_key.drop(columns=[key], inplace=True) results.append(result_key) merged_results = pd.concat(results, axis=1) merged_results["score"] = merged_results.mean(axis=1, skipna=False) merged_results.sort_values("score", ascending=False, inplace=True) return merged_results naive_retriever = NaiveRetriever(df.set_index("name")) Let's use the first query from our generated list above, and search for monsters using our naive_retriever: query = { 'look': 'glowing', 'habitat': 'dark places', 'behavior': 'light manipulation' } naive_retriever.search(query) Our naive_retriever returns the following search results for each attribute: id score_look look Whispering Shade 0.503578 Shadowy, amorphous being with glowing eyes Sandstorm Djinn 0.407344 Swirling vortex of sand with glowing symbols Luminoth 0.378619 Moth-like creature with glowing wings and antenna Awesome! Our returned monster results are relevant - they all have some "glowing" characteristic. Let's see what the naive approach returns when we search the other two attributes. id score_habitat habitat Whispering Shade 0.609567 Dark forests and abandoned buildings Fungal Network 0.438856 Underground caverns and damp forests Thornvine Elemental 0.423421 Overgrown ruins and dense jungles id score_behavior behavior Living Graffiti 0.385741 Shapeshifts to blend with surroundings and absorbs pigments Crystalwing Drake 0.385211 Hoards precious gems and can refract light into powerful beams Luminoth 0.345566 Emits soothing light patterns to communicate and attract prey All the retrieved monsters do possess the wanted attributes. At first glance, the naive search results may seem promising. But we need to find monsters that possess all three attributes simultaneously. Let's merge our results to see how well our monsters do at achieving this goal: id score_look score_habitat score_behavior Whispering Shade 0.503578 0.609567 Sandstorm Djinn 0.407344 Luminoth 0.378619 0.345566 Fungal Network 0.438856 Thornvine Elemental 0.423421 Living Graffiti 0.385741 Crystalwing Drake 0.385211 And here, the limits of the naive approach become obvious. Let's evaluate: Relevance by attribute: look: Three monsters were retrieved (Whispering Shade, Sandstorm Djinn, and Luminoth). habitat: Only one monster from the look results was relevant (Whispering Shade). behavior: Only one monster from the look results was relevant (Luminoth), but it's different from the one relevant for habitat. Overall relevance: No single monster was retrieved for all three attributes simultaneously. The results are fragmented: different monsters are relevant for different attributes. In short, the naive search approach fails to find monsters that satisfy all criteria at once. Maybe we can fix this issue by proactively retrieving more monsters for each attribute? Let's try it with 6 monsters per attribute, instead of 3. Let's take a look at what this approach generates: id score_look score_habitat score_behavior Whispering Shade 0.503578 0.609567 Sandstorm Djinn 0.407344 0.365061 Luminoth 0.378619 0.345566 Nebula Jellyfish 0.36627 0.259969 Dreamweaver Octopus 0.315679 Quantum Firefly 0.288578 Fungal Network 0.438856 Thornvine Elemental 0.423421 Mist Phantom 0.366816 0.236649 Stoneheart Golem 0.342287 Living Graffiti 0.385741 Crystalwing Drake 0.385211 Aqua Wraith 0.283581 We've now retrieved 13 monsters (more than half of our tiny dataset!), and still have the same issue: not one of these monsters was retrieved for all three attributes. Increasing the number of retrieved monsters (beyond 6) might solve our problem, but it creates additional issues: In production, retrieving more results (multiple kNN searches) lengthens search time noticeably. For each new attribute we introduce, our chances of finding a "complete" monster - with all the attributes in our query - drops exponentially. To prevent this, we have to retrieve many more nearest neighbors (monsters), making the total number of retrieved monsters grow exponentially. We still have no guarantee we'll retrieve monsters that possess all our desired attributes. If we do manage to retrieve monsters that satisfy all criteria at once, we'll have to expend additional overhead reconciling results. In sum, the naive approach is too uncertain and inefficient for viable multi-attribute search, especially in production. Superlinked approach Let's implement our second approach to see if it does any better than the naive one. First, we define the schema, spaces, index, and query: @schema class Monster: id: IdField look: String habitat: String behavior: String monster = Monster() look_space = TextSimilaritySpace(text=monster.look, model=MODEL_NAME) habitat_space = TextSimilaritySpace(text=monster.habitat, model=MODEL_NAME) behavior_space = TextSimilaritySpace(text=monster.behavior, model=MODEL_NAME) monster_index = Index([look_space, habitat_space, behavior_space]) monster_query = ( Query( monster_index, weights={ look_space: Param("look_weight"), habitat_space: Param("habitat_weight"), behavior_space: Param("behavior_weight"), }, ) .find(monster) .similar(look_space.text, Param("look")) .similar(habitat_space.text, Param("habitat")) .similar(behavior_space.text, Param("behavior")) .limit(LIMIT) ) default_weights = { "look_weight": 1.0, "habitat_weight": 1.0, "behavior_weight": 1.0 } Now, we start the executor and upload the data: monster_parser = DataFrameParser(monster, mapping={monster.id: "name"}) source: InMemorySource = InMemorySource(monster, parser=monster_parser) executor = InMemoryExecutor(sources=[source], indices=[monster_index]) app = executor.run() source.put([df]) Let's run the same query we ran in our naive approach implementation above: query = { 'look': 'glowing', 'habitat': 'dark places', 'behavior': 'light manipulation' } app.query( monster_query, limit=LIMIT, **query, **default_weights ) id score look habitat behavior Whispering Shade 0.376738 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Luminoth 0.340084 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Living Graffiti 0.330587 Two-dimensional, colorful creature that inhabits flat surfaces Urban areas, particularly walls and billboards Shapeshifts to blend with surroundings and absorbs pigments Et voila! This time, each of our top returned monsters ranks highly in a score that represents a kind of "mean" of all three characteristics we want our monster to have. Let's break each monster's score out in detail: id look habitat behavior total Whispering Shade 0.167859 0.203189 0.005689 0.376738 Luminoth 0.126206 0.098689 0.115189 0.340084 Living Graffiti 0.091063 0.110944 0.12858 0.330587 Our second and third results, Luminoth and Living Graffiti, both possess all three of the desired characteristics. The top result, Whispering Shade, though it's less relevant in terms of light manipulation - as reflected in its behavior score (0.006), has "glowing" features and a dark environment that make its look (0.168) and habitat (0.203) scores very high, giving it the highest total score (0.377), making it the most relevant monster overall. What an improvement! Can we can replicate our results? Let's try another query and find out. query = { 'look': 'shapeshifting', 'habitat': 'varied landscapes', 'behavior': 'illusion creation' } id score look habitat behavior Mist Phantom 0.489574 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Zephyr Dancer 0.342075 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Whispering Shade 0.337434 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Great! Our outcomes are excellent again. What if we want to find monsters that are similar to a specific monster from our dataset? Let's try it with a monster we haven't seen yet - Harmonic Coral. We could extract attributes for this monster and create query parameters manually. But Superlinked has a with_vector method we can use on the query object. Because each monster's id is its name, we can express our request as simply as: app.query( monster_query.with_vector(monster, "Harmonic Coral"), **default_weights, limit=LIMIT ) id score look habitat behavior Harmonic Coral 1 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.402288 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Aqua Wraith 0.330869 Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents The top result is the most relevant one, Harmonic Coral itself, as expected. The other two monsters our search retrieves are Dreamweaver Octopus and Aqua Wraith. Both share important thematic (attribute) elements with Harmonic Coral: Aquatic habitats (habitat) Ability to influence or manipulate their environment (behavior) Dynamic or fluid visual characteristics (look) Attribute weighting Suppose, now, that we want to give more importance to the look attribute. The Superlinked framework lets us easily adjust weights at query time. For easy comparison, we'll search for monsters similar to Harmonic Coral, but with our weights adjusted to favor look. weights = { "look_weight": 1.0, "habitat_weight": 0, "behavior_weight": 0 } app.query( monster_query.with_vector(monster, "Harmonic Coral"), limit=LIMIT, **weights ) id score look habitat behavior Harmonic Coral 0.57735 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Thornvine Elemental 0.252593 Plant-like creature with a body of twisted vines and thorns Overgrown ruins and dense jungles Rapidly grows and controls surrounding plant life Plasma Serpent 0.243241 Snake-like creature made of crackling energy Electrical storms and power plants Feeds on electrical currents and can short-circuit technology Our results all (appropriately) have similar appearances - "Branching with vibrating tendrils", "Plant-like creature with a body of twisted vines and thorns", "Snake-like". Now, let's do another search, ignoring appearance, and looking instead for monsters that are similar in terms of habitat and behavior simultaneously: weights = { "look_weight": 0, "habitat_weight": 1.0, "behavior_weight": 1.0 } id score look habitat behavior Harmonic Coral 0.816497 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.357656 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Mist Phantom 0.288106 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Again, the Superlinked approach produces great results. All three monsters live in watery environments and possess mind-controlling abilities. Finally, let's try another search, weighting all three attributes differently - to find monsters that in comparison to Harmonic Coral look somewhat similar, live in very different habitats, and possess very similar behavior: weights = { "look_weight": 0.5, "habitat_weight": -1.0, "behavior_weight": 1.0 } id score look habitat behavior Harmonic Coral 0.19245 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Luminoth 0.149196 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Zephyr Dancer 0.136456 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Great results again! Our two other retrieved monsters — Luminoth and Zephyr Dancer — have behavior similar to Harmonic Coral and live in habitats different from Harmonic Coral's. They also look very different from Harmonic Coral. (While Harmonic Coral's tendrils and Luminoth's antenna are somewhat similar features, we only down-weighted the look_weight by 0.5, and the resemblance between the two monsters ends there.) Let's see how these monsters' overall scores break out in terms of individual attributes: id look habitat behavior total Harmonic Coral 0.19245 -0.3849 0.3849 0.19245 Luminoth 0.052457 -0.068144 0.164884 0.149196 Zephyr Dancer 0.050741 -0.079734 0.165449 0.136456 By negatively weighting habitat_weight (-1.0), we deliberately "push away" monsters with similar habitats and instead surface monsters whose environments are different from Harmonic Coral's - as seen in Luminoth's and Zephyr Dancer's negative habitat scores. Luminoth's and Zephyr Dancer's behavior scores are relatively high, indicating their behavioral similarity to Harmonic Coral. Their look scores are positive but lower, reflecting some but not extreme visual similarity to Harmonic Coral. In short, our strategy of downweighting habitat_weight to -1.0 and look_weight to 0.5 but keeping behavior_weight at 1.0 proves effective in surfacing monsters that share key behavioral characteristics with Harmonic Coral but have very different environments and look at least somewhat different. Conclusion Multi-attribute vector search is a significant advance in information retrieval, offering more accuracy, contextual understanding, and flexibility than basic semantic similarity search. Still, our naive approach (above) - storing and searching attribute vectors separately, then combining results - is limited in ability, subtlety, and efficiency when we need to retrieve objects with multiple simultaneous attributes. (Moreover, multiple kNN searches take more time than a single search with concatenated vectors.) To handle scenarios like this, it's better to store all your attribute vectors in the same vector store and perform a single search, weighting your attributes at query time. The Superlinked approach is more accurate, efficient, and scalable than the naive approach for any application that requires fast, reliable, nuanced, multi-attribute vector retrieval - whether your use case is tackling real world data challenges in your e-commerce or recommendation system... or something entirely different, like battling monsters. Contributors Andrey Pikunov, Author Mór Kapronczay, Editor Originally published here. An AI-powered Army of Darkness It's game night, your friends are perched around the games table, waiting to see what Dungeons & Dragons (D&D) character they'll become and quest they'll embark on. Tonight, you're Dungeon Master (storyteller and guide), crafter of thrilling encounters to challenge and enthrall your players. Your trusty D&D Monster Manual contains thousands of creatures. Finding the perfect monster for each situation among the myriad options can be overwhelming. The ideal foe needs to match the setting, difficulty, and narrative of the moment. thousands What if we could create a tool that instantly finds the monster most suited to each scenario? A tool that considers multiple factors simultaneously , ensuring each encounter is as immersive and exciting as possible? tool that considers multiple factors simultaneously Let's embark on a quest of our own: build the ultimate monster-finding system, using the power of multi-attribute vector search! Creating creatures with vector search, why do it? Vector search represents a revolution in information retrieval. Vector embedding - by taking account of context and semantic meaning - empowers vector search to return more relevant and accurate results, handle not just structured but also unstructured data and multiple languages, and scale. But to generate high quality responses in real-world applications, we often need to assign different weights to specific attributes of our data objects. Vector search There are two common approaches to multi-attribute vector search. Both start by separately embedding each attribute of a data object. The main difference between these two approaches is in how our embeddings are stored and searched . stored searched the naive approach - store each attribute vector in separate vector stores (one per attribute), perform a separate search for each attribute, combine search results, and post-process (e.g., weight) as required. the Superlinked approach - concatenate and store all attribute vectors in the same vector store (using Superlinked's built-in funtionality), which allows us to search just once, with attendant efficiency gains. Superlinked's spaces also let us weight each attribute at query time to surface more relevant results, with no post-processing. the naive approach - store each attribute vector in separate vector stores (one per attribute), perform a separate search for each attribute, combine search results, and post-process (e.g., weight) as required. naive the Superlinked approach - concatenate and store all attribute vectors in the same vector store (using Superlinked's built-in funtionality), which allows us to search just once , with attendant efficiency gains. Superlinked's spaces also let us weight each attribute at query time to surface more relevant results, with no post-processing. Superlinked Superlinked search just once spaces also Below, we'll use these two approaches to implement a multi-attribute vector search tool - a Dungeons and Dragons monster finder! Our simple implementations, especially the second, will illustrate how to create more powerful and flexible search systems, ones that can handle complex, multi-faceted queries with ease, whatever your use case. If you're new to vector similarity search, don't worry! We've got you covered - check out our building blocks articles . building blocks articles Okay, let's go monster hunting! Dataset First, we'll generate a small synthetic dataset of monsters, by prompting a Large Language Model (LLM): Generate two JSON lists: 'monsters' and 'queries'. 1. 'monsters' list: Create 20 unique monsters with the following properties: - name: A distinctive name - look: Brief description of appearance (2-3 sentences) - habitat: Where the monster lives (2-3 sentences) - behavior: How the monster acts (2-3 sentences) Ensure some monsters share similar features while remaining distinct. 2. 'queries' list: Create 5 queries to search for monsters: - Each query should be in the format: {look: "...", habitat: "...", behavior: "..."} - Use simple, brief descriptions (1-3 words per field) - Make queries somewhat general to match multiple monsters Output format: { "monsters": [ {"name": "...", "look": "...", "habitat": "...", "behavior": "..."}, ... ], "queries": [ {"look": "...", "habitat": "...", "behavior": "..."}, ... ] } Generate two JSON lists: 'monsters' and 'queries'. 1. 'monsters' list: Create 20 unique monsters with the following properties: - name: A distinctive name - look: Brief description of appearance (2-3 sentences) - habitat: Where the monster lives (2-3 sentences) - behavior: How the monster acts (2-3 sentences) Ensure some monsters share similar features while remaining distinct. 2. 'queries' list: Create 5 queries to search for monsters: - Each query should be in the format: {look: "...", habitat: "...", behavior: "..."} - Use simple, brief descriptions (1-3 words per field) - Make queries somewhat general to match multiple monsters Output format: { "monsters": [ {"name": "...", "look": "...", "habitat": "...", "behavior": "..."}, ... ], "queries": [ {"look": "...", "habitat": "...", "behavior": "..."}, ... ] } Let's take a look at a sample of the dataset our LLM generated. Note: LLM generation is non-deterministic, so your results may differ. Here are our first five monsters: # name look habitat behavior 0 Luminoth Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey 1 Aqua Wraith Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents 2 Stoneheart Golem Massive humanoid composed of interlocking rock formations Rocky mountains and ancient ruins Hibernates for centuries, awakens to protect its territory 3 Whispering Shade Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths 4 Zephyr Dancer Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates # name look habitat behavior 0 Luminoth Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey 1 Aqua Wraith Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents 2 Stoneheart Golem Massive humanoid composed of interlocking rock formations Rocky mountains and ancient ruins Hibernates for centuries, awakens to protect its territory 3 Whispering Shade Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths 4 Zephyr Dancer Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates # name look habitat behavior # # name name look look habitat habitat behavior behavior 0 Luminoth Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey 0 0 Luminoth Luminoth Moth-like creature with glowing wings and antenna Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Emits soothing light patterns to communicate and attract prey 1 Aqua Wraith Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents 1 1 Aqua Wraith Aqua Wraith Translucent humanoid figure made of flowing water Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents Shapeshifts to blend with water bodies and controls currents 2 Stoneheart Golem Massive humanoid composed of interlocking rock formations Rocky mountains and ancient ruins Hibernates for centuries, awakens to protect its territory 2 2 Stoneheart Golem Stoneheart Golem Massive humanoid composed of interlocking rock formations Massive humanoid composed of interlocking rock formations Rocky mountains and ancient ruins Rocky mountains and ancient ruins Hibernates for centuries, awakens to protect its territory Hibernates for centuries, awakens to protect its territory 3 Whispering Shade Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths 3 3 Whispering Shade Whispering Shade Shadowy, amorphous being with glowing eyes Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Feeds on fear and whispers unsettling truths 4 Zephyr Dancer Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates 4 4 Zephyr Dancer Zephyr Dancer Graceful avian creature with iridescent feathers Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Creates mesmerizing aerial displays to attract mates ...and our generated queries: Look Habitat Behavior 0 Glowing Dark places Light manipulation 1 Elemental Extreme environments Environmental control 2 Shapeshifting Varied landscapes Illusion creation 3 Crystalline Mineral-rich areas Energy absorption 4 Ethereal Atmospheric Mind influence Look Habitat Behavior 0 Glowing Dark places Light manipulation 1 Elemental Extreme environments Environmental control 2 Shapeshifting Varied landscapes Illusion creation 3 Crystalline Mineral-rich areas Energy absorption 4 Ethereal Atmospheric Mind influence Look Habitat Behavior Look Look Look Habitat Habitat Habitat Behavior Behavior Behavior 0 Glowing Dark places Light manipulation 0 0 Glowing Glowing Dark places Dark places Light manipulation Light manipulation 1 Elemental Extreme environments Environmental control 1 1 Elemental Elemental Extreme environments Extreme environments Environmental control Environmental control 2 Shapeshifting Varied landscapes Illusion creation 2 2 Shapeshifting Shapeshifting Varied landscapes Varied landscapes Illusion creation Illusion creation 3 Crystalline Mineral-rich areas Energy absorption 3 3 Crystalline Crystalline Mineral-rich areas Mineral-rich areas Energy absorption Energy absorption 4 Ethereal Atmospheric Mind influence 4 4 Ethereal Ethereal Atmospheric Atmospheric Mind influence Mind influence See original dataset and query examples here . here Retrieval Let's set up parameters we'll use in both of our approaches - naive and Superlinked - below. We generate our vector embeddings with: sentence-transformers/all-mpnet-base-v2. sentence-transformers/all-mpnet-base-v2. For simplicity's sake, we'll limit our output to the top 3 matches. (For complete code, including necessary imports and helper functions, see the notebook .) notebook LIMIT = 3 MODEL_NAME = "sentence-transformers/all-mpnet-base-v2" LIMIT = 3 MODEL_NAME = "sentence-transformers/all-mpnet-base-v2" Now, let's get our multi-attribute monster search under way! First, we'll try the naive approach. naive Naive approach In our naive approach, we embed attributes independently and store them in different indices. At query time, we run multiple kNN-searches on all the indices, and then combine all our partial results into one. We start by defining a class NaiveRetriever NaiveRetriever to perform similarity-based search on our dataset, using our all-mpnet-base-v2 -generated embeddings. all-mpnet-base-v2 class NaiveRetriever: def __init__(self, data: pd.DataFrame): self.model = SentenceTransformer(MODEL_NAME) self.data = data.copy() self.ids = self.data.index.to_list() self.knns = {} for key in self.data: embeddings = self.model.encode(self.data[key].values) knn = NearestNeighbors(metric="cosine").fit(embeddings) self.knns[key] = knn def search_key(self, key: str, value: str, limit: int = LIMIT) -> pd.DataFrame: embedding = self.model.encode(value) knn = self.knns[key] distances, indices = knn.kneighbors( [embedding], n_neighbors=limit, return_distance=True ) ids = [self.ids[i] for i in indices[0]] similarities = (1 - distances).flatten() # by definition: # cosine distance = 1 - cosine similarity result = pd.DataFrame( {"id": ids, f"score_{key}": similarities, key: self.data[key][ids]} ) result.set_index("id", inplace=True) return result def search(self, query: dict, limit: int = LIMIT) -> pd.DataFrame: results = [] for key, value in query.items(): if key not in self.knns: continue result_key = self.search_key(key, value, limit=limit) result_key.drop(columns=[key], inplace=True) results.append(result_key) merged_results = pd.concat(results, axis=1) merged_results["score"] = merged_results.mean(axis=1, skipna=False) merged_results.sort_values("score", ascending=False, inplace=True) return merged_results naive_retriever = NaiveRetriever(df.set_index("name")) class NaiveRetriever: def __init__(self, data: pd.DataFrame): self.model = SentenceTransformer(MODEL_NAME) self.data = data.copy() self.ids = self.data.index.to_list() self.knns = {} for key in self.data: embeddings = self.model.encode(self.data[key].values) knn = NearestNeighbors(metric="cosine").fit(embeddings) self.knns[key] = knn def search_key(self, key: str, value: str, limit: int = LIMIT) -> pd.DataFrame: embedding = self.model.encode(value) knn = self.knns[key] distances, indices = knn.kneighbors( [embedding], n_neighbors=limit, return_distance=True ) ids = [self.ids[i] for i in indices[0]] similarities = (1 - distances).flatten() # by definition: # cosine distance = 1 - cosine similarity result = pd.DataFrame( {"id": ids, f"score_{key}": similarities, key: self.data[key][ids]} ) result.set_index("id", inplace=True) return result def search(self, query: dict, limit: int = LIMIT) -> pd.DataFrame: results = [] for key, value in query.items(): if key not in self.knns: continue result_key = self.search_key(key, value, limit=limit) result_key.drop(columns=[key], inplace=True) results.append(result_key) merged_results = pd.concat(results, axis=1) merged_results["score"] = merged_results.mean(axis=1, skipna=False) merged_results.sort_values("score", ascending=False, inplace=True) return merged_results naive_retriever = NaiveRetriever(df.set_index("name")) Let's use the first query from our generated list above, and search for monsters using our naive_retriever : naive_retriever query = { 'look': 'glowing', 'habitat': 'dark places', 'behavior': 'light manipulation' } naive_retriever.search(query) query = { 'look': 'glowing', 'habitat': 'dark places', 'behavior': 'light manipulation' } naive_retriever.search(query) Our naive_retriever naive_retriever returns the following search results for each attribute: id score_look look Whispering Shade 0.503578 Shadowy, amorphous being with glowing eyes Sandstorm Djinn 0.407344 Swirling vortex of sand with glowing symbols Luminoth 0.378619 Moth-like creature with glowing wings and antenna id score_look look Whispering Shade 0.503578 Shadowy, amorphous being with glowing eyes Sandstorm Djinn 0.407344 Swirling vortex of sand with glowing symbols Luminoth 0.378619 Moth-like creature with glowing wings and antenna id score_look look id id score_look score_look look look Whispering Shade 0.503578 Shadowy, amorphous being with glowing eyes Whispering Shade Whispering Shade 0.503578 0.503578 Shadowy, amorphous being with glowing eyes Shadowy, amorphous being with glowing eyes Sandstorm Djinn 0.407344 Swirling vortex of sand with glowing symbols Sandstorm Djinn Sandstorm Djinn 0.407344 0.407344 Swirling vortex of sand with glowing symbols Swirling vortex of sand with glowing symbols Luminoth 0.378619 Moth-like creature with glowing wings and antenna Luminoth Luminoth 0.378619 0.378619 Moth-like creature with glowing wings and antenna Moth-like creature with glowing wings and antenna Awesome! Our returned monster results are relevant - they all have some "glowing" characteristic. Let's see what the naive approach returns when we search the other two attributes. id score_habitat habitat Whispering Shade 0.609567 Dark forests and abandoned buildings Fungal Network 0.438856 Underground caverns and damp forests Thornvine Elemental 0.423421 Overgrown ruins and dense jungles id score_habitat habitat Whispering Shade 0.609567 Dark forests and abandoned buildings Fungal Network 0.438856 Underground caverns and damp forests Thornvine Elemental 0.423421 Overgrown ruins and dense jungles id score_habitat habitat id id id score_habitat score_habitat score_habitat habitat habitat habitat Whispering Shade 0.609567 Dark forests and abandoned buildings Whispering Shade Whispering Shade 0.609567 0.609567 Dark forests and abandoned buildings Dark forests and abandoned buildings Fungal Network 0.438856 Underground caverns and damp forests Fungal Network Fungal Network 0.438856 0.438856 Underground caverns and damp forests Underground caverns and damp forests Thornvine Elemental 0.423421 Overgrown ruins and dense jungles Thornvine Elemental Thornvine Elemental 0.423421 0.423421 Overgrown ruins and dense jungles Overgrown ruins and dense jungles id score_behavior behavior Living Graffiti 0.385741 Shapeshifts to blend with surroundings and absorbs pigments Crystalwing Drake 0.385211 Hoards precious gems and can refract light into powerful beams Luminoth 0.345566 Emits soothing light patterns to communicate and attract prey id score_behavior behavior Living Graffiti 0.385741 Shapeshifts to blend with surroundings and absorbs pigments Crystalwing Drake 0.385211 Hoards precious gems and can refract light into powerful beams Luminoth 0.345566 Emits soothing light patterns to communicate and attract prey id score_behavior behavior id id id score_behavior score_behavior score_behavior behavior behavior behavior Living Graffiti 0.385741 Shapeshifts to blend with surroundings and absorbs pigments Living Graffiti Living Graffiti 0.385741 0.385741 Shapeshifts to blend with surroundings and absorbs pigments Shapeshifts to blend with surroundings and absorbs pigments Crystalwing Drake 0.385211 Hoards precious gems and can refract light into powerful beams Crystalwing Drake Crystalwing Drake 0.385211 0.385211 Hoards precious gems and can refract light into powerful beams Hoards precious gems and can refract light into powerful beams Luminoth 0.345566 Emits soothing light patterns to communicate and attract prey Luminoth Luminoth 0.345566 0.345566 Emits soothing light patterns to communicate and attract prey Emits soothing light patterns to communicate and attract prey All the retrieved monsters do possess the wanted attributes. At first glance, the naive search results may seem promising. But we need to find monsters that possess all three attributes simultaneously . Let's merge our results to see how well our monsters do at achieving this goal: all three attributes simultaneously id score_look score_habitat score_behavior Whispering Shade 0.503578 0.609567 Sandstorm Djinn 0.407344 Luminoth 0.378619 0.345566 Fungal Network 0.438856 Thornvine Elemental 0.423421 Living Graffiti 0.385741 Crystalwing Drake 0.385211 id score_look score_habitat score_behavior Whispering Shade 0.503578 0.609567 Sandstorm Djinn 0.407344 Luminoth 0.378619 0.345566 Fungal Network 0.438856 Thornvine Elemental 0.423421 Living Graffiti 0.385741 Crystalwing Drake 0.385211 id score_look score_habitat score_behavior id id id score_look score_look score_look score_habitat score_habitat score_habitat score_behavior score_behavior score_behavior Whispering Shade 0.503578 0.609567 Whispering Shade Whispering Shade 0.503578 0.503578 0.609567 0.609567 Sandstorm Djinn 0.407344 Sandstorm Djinn Sandstorm Djinn 0.407344 0.407344 Luminoth 0.378619 0.345566 Luminoth Luminoth 0.378619 0.378619 0.345566 0.345566 Fungal Network 0.438856 Fungal Network Fungal Network 0.438856 0.438856 Thornvine Elemental 0.423421 Thornvine Elemental Thornvine Elemental 0.423421 0.423421 Living Graffiti 0.385741 Living Graffiti Living Graffiti 0.385741 0.385741 Crystalwing Drake 0.385211 Crystalwing Drake Crystalwing Drake 0.385211 0.385211 And here, the limits of the naive approach become obvious. Let's evaluate: Relevance by attribute: look: Three monsters were retrieved (Whispering Shade, Sandstorm Djinn, and Luminoth). habitat: Only one monster from the look results was relevant (Whispering Shade). behavior: Only one monster from the look results was relevant (Luminoth), but it's different from the one relevant for habitat. Overall relevance: No single monster was retrieved for all three attributes simultaneously. The results are fragmented: different monsters are relevant for different attributes. Relevance by attribute: look: Three monsters were retrieved (Whispering Shade, Sandstorm Djinn, and Luminoth). habitat: Only one monster from the look results was relevant (Whispering Shade). behavior: Only one monster from the look results was relevant (Luminoth), but it's different from the one relevant for habitat. look: Three monsters were retrieved (Whispering Shade, Sandstorm Djinn, and Luminoth). habitat: Only one monster from the look results was relevant (Whispering Shade). behavior: Only one monster from the look results was relevant (Luminoth), but it's different from the one relevant for habitat. look : Three monsters were retrieved (Whispering Shade, Sandstorm Djinn, and Luminoth). look habitat : Only one monster from the look results was relevant (Whispering Shade). habitat look behavior : Only one monster from the look results was relevant (Luminoth), but it's different from the one relevant for habitat . behavior look habitat Overall relevance: No single monster was retrieved for all three attributes simultaneously. The results are fragmented: different monsters are relevant for different attributes. No single monster was retrieved for all three attributes simultaneously. The results are fragmented: different monsters are relevant for different attributes. No single monster was retrieved for all three attributes simultaneously. The results are fragmented: different monsters are relevant for different attributes. In short, the naive search approach fails to find monsters that satisfy all criteria at once. Maybe we can fix this issue by proactively retrieving more monsters for each attribute? Let's try it with 6 monsters per attribute, instead of 3. Let's take a look at what this approach generates: Maybe we can fix this issue by proactively retrieving more monsters for each attribute? id score_look score_habitat score_behavior Whispering Shade 0.503578 0.609567 Sandstorm Djinn 0.407344 0.365061 Luminoth 0.378619 0.345566 Nebula Jellyfish 0.36627 0.259969 Dreamweaver Octopus 0.315679 Quantum Firefly 0.288578 Fungal Network 0.438856 Thornvine Elemental 0.423421 Mist Phantom 0.366816 0.236649 Stoneheart Golem 0.342287 Living Graffiti 0.385741 Crystalwing Drake 0.385211 Aqua Wraith 0.283581 id score_look score_habitat score_behavior Whispering Shade 0.503578 0.609567 Sandstorm Djinn 0.407344 0.365061 Luminoth 0.378619 0.345566 Nebula Jellyfish 0.36627 0.259969 Dreamweaver Octopus 0.315679 Quantum Firefly 0.288578 Fungal Network 0.438856 Thornvine Elemental 0.423421 Mist Phantom 0.366816 0.236649 Stoneheart Golem 0.342287 Living Graffiti 0.385741 Crystalwing Drake 0.385211 Aqua Wraith 0.283581 id score_look score_habitat score_behavior id id id score_look score_look score_look score_habitat score_habitat score_habitat score_behavior score_behavior score_behavior Whispering Shade 0.503578 0.609567 Whispering Shade Whispering Shade 0.503578 0.503578 0.609567 0.609567 Sandstorm Djinn 0.407344 0.365061 Sandstorm Djinn Sandstorm Djinn 0.407344 0.407344 0.365061 0.365061 Luminoth 0.378619 0.345566 Luminoth Luminoth 0.378619 0.378619 0.345566 0.345566 Nebula Jellyfish 0.36627 0.259969 Nebula Jellyfish Nebula Jellyfish 0.36627 0.36627 0.259969 0.259969 Dreamweaver Octopus 0.315679 Dreamweaver Octopus Dreamweaver Octopus 0.315679 0.315679 Quantum Firefly 0.288578 Quantum Firefly Quantum Firefly 0.288578 0.288578 Fungal Network 0.438856 Fungal Network Fungal Network 0.438856 0.438856 Thornvine Elemental 0.423421 Thornvine Elemental Thornvine Elemental 0.423421 0.423421 Mist Phantom 0.366816 0.236649 Mist Phantom Mist Phantom 0.366816 0.366816 0.236649 0.236649 Stoneheart Golem 0.342287 Stoneheart Golem Stoneheart Golem 0.342287 0.342287 Living Graffiti 0.385741 Living Graffiti Living Graffiti 0.385741 0.385741 Crystalwing Drake 0.385211 Crystalwing Drake Crystalwing Drake 0.385211 0.385211 Aqua Wraith 0.283581 Aqua Wraith Aqua Wraith 0.283581 0.283581 We've now retrieved 13 monsters (more than half of our tiny dataset!), and still have the same issue: not one of these monsters was retrieved for all three attributes. still Increasing the number of retrieved monsters (beyond 6) might solve our problem, but it creates additional issues: might In production, retrieving more results (multiple kNN searches) lengthens search time noticeably. For each new attribute we introduce, our chances of finding a "complete" monster - with all the attributes in our query - drops exponentially. To prevent this, we have to retrieve many more nearest neighbors (monsters), making the total number of retrieved monsters grow exponentially. We still have no guarantee we'll retrieve monsters that possess all our desired attributes. If we do manage to retrieve monsters that satisfy all criteria at once, we'll have to expend additional overhead reconciling results. In production, retrieving more results (multiple kNN searches) lengthens search time noticeably. For each new attribute we introduce, our chances of finding a "complete" monster - with all the attributes in our query - drops exponentially. To prevent this, we have to retrieve many more nearest neighbors (monsters), making the total number of retrieved monsters grow exponentially. We still have no guarantee we'll retrieve monsters that possess all our desired attributes. If we do manage to retrieve monsters that satisfy all criteria at once, we'll have to expend additional overhead reconciling results. In sum, the naive approach is too uncertain and inefficient for viable multi-attribute search, especially in production. Superlinked approach Let's implement our second approach to see if it does any better than the naive one. First, we define the schema, spaces, index, and query: @schema class Monster: id: IdField look: String habitat: String behavior: String monster = Monster() look_space = TextSimilaritySpace(text=monster.look, model=MODEL_NAME) habitat_space = TextSimilaritySpace(text=monster.habitat, model=MODEL_NAME) behavior_space = TextSimilaritySpace(text=monster.behavior, model=MODEL_NAME) monster_index = Index([look_space, habitat_space, behavior_space]) monster_query = ( Query( monster_index, weights={ look_space: Param("look_weight"), habitat_space: Param("habitat_weight"), behavior_space: Param("behavior_weight"), }, ) .find(monster) .similar(look_space.text, Param("look")) .similar(habitat_space.text, Param("habitat")) .similar(behavior_space.text, Param("behavior")) .limit(LIMIT) ) default_weights = { "look_weight": 1.0, "habitat_weight": 1.0, "behavior_weight": 1.0 } @schema class Monster: id: IdField look: String habitat: String behavior: String monster = Monster() look_space = TextSimilaritySpace(text=monster.look, model=MODEL_NAME) habitat_space = TextSimilaritySpace(text=monster.habitat, model=MODEL_NAME) behavior_space = TextSimilaritySpace(text=monster.behavior, model=MODEL_NAME) monster_index = Index([look_space, habitat_space, behavior_space]) monster_query = ( Query( monster_index, weights={ look_space: Param("look_weight"), habitat_space: Param("habitat_weight"), behavior_space: Param("behavior_weight"), }, ) .find(monster) .similar(look_space.text, Param("look")) .similar(habitat_space.text, Param("habitat")) .similar(behavior_space.text, Param("behavior")) .limit(LIMIT) ) default_weights = { "look_weight": 1.0, "habitat_weight": 1.0, "behavior_weight": 1.0 } Now, we start the executor and upload the data: monster_parser = DataFrameParser(monster, mapping={monster.id: "name"}) source: InMemorySource = InMemorySource(monster, parser=monster_parser) executor = InMemoryExecutor(sources=[source], indices=[monster_index]) app = executor.run() source.put([df]) monster_parser = DataFrameParser(monster, mapping={monster.id: "name"}) source: InMemorySource = InMemorySource(monster, parser=monster_parser) executor = InMemoryExecutor(sources=[source], indices=[monster_index]) app = executor.run() source.put([df]) Let's run the same query we ran in our naive approach implementation above: query = { 'look': 'glowing', 'habitat': 'dark places', 'behavior': 'light manipulation' } app.query( monster_query, limit=LIMIT, **query, **default_weights ) query = { 'look': 'glowing', 'habitat': 'dark places', 'behavior': 'light manipulation' } app.query( monster_query, limit=LIMIT, **query, **default_weights ) id score look habitat behavior Whispering Shade 0.376738 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Luminoth 0.340084 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Living Graffiti 0.330587 Two-dimensional, colorful creature that inhabits flat surfaces Urban areas, particularly walls and billboards Shapeshifts to blend with surroundings and absorbs pigments id score look habitat behavior Whispering Shade 0.376738 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Luminoth 0.340084 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Living Graffiti 0.330587 Two-dimensional, colorful creature that inhabits flat surfaces Urban areas, particularly walls and billboards Shapeshifts to blend with surroundings and absorbs pigments id score look habitat behavior id id id score score score look look look habitat habitat habitat behavior behavior behavior Whispering Shade 0.376738 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Whispering Shade Whispering Shade 0.376738 0.376738 Shadowy, amorphous being with glowing eyes Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Feeds on fear and whispers unsettling truths Luminoth 0.340084 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Luminoth Luminoth 0.340084 0.340084 Moth-like creature with glowing wings and antenna Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Emits soothing light patterns to communicate and attract prey Living Graffiti 0.330587 Two-dimensional, colorful creature that inhabits flat surfaces Urban areas, particularly walls and billboards Shapeshifts to blend with surroundings and absorbs pigments Living Graffiti Living Graffiti 0.330587 0.330587 Two-dimensional, colorful creature that inhabits flat surfaces Two-dimensional, colorful creature that inhabits flat surfaces Urban areas, particularly walls and billboards Urban areas, particularly walls and billboards Shapeshifts to blend with surroundings and absorbs pigments Shapeshifts to blend with surroundings and absorbs pigments Et voila! This time, each of our top returned monsters ranks highly in a score that represents a kind of "mean" of all three characteristics we want our monster to have. Let's break each monster's score out in detail: id look habitat behavior total Whispering Shade 0.167859 0.203189 0.005689 0.376738 Luminoth 0.126206 0.098689 0.115189 0.340084 Living Graffiti 0.091063 0.110944 0.12858 0.330587 id look habitat behavior total Whispering Shade 0.167859 0.203189 0.005689 0.376738 Luminoth 0.126206 0.098689 0.115189 0.340084 Living Graffiti 0.091063 0.110944 0.12858 0.330587 id look habitat behavior total id id id look look look habitat habitat habitat behavior behavior behavior total total total Whispering Shade 0.167859 0.203189 0.005689 0.376738 Whispering Shade Whispering Shade 0.167859 0.167859 0.203189 0.203189 0.005689 0.005689 0.376738 0.376738 Luminoth 0.126206 0.098689 0.115189 0.340084 Luminoth Luminoth 0.126206 0.126206 0.098689 0.098689 0.115189 0.115189 0.340084 0.340084 Living Graffiti 0.091063 0.110944 0.12858 0.330587 Living Graffiti Living Graffiti 0.091063 0.091063 0.110944 0.110944 0.12858 0.12858 0.330587 0.330587 Our second and third results, Luminoth and Living Graffiti, both possess all three of the desired characteristics. The top result, Whispering Shade, though it's less relevant in terms of light manipulation - as reflected in its behavior score (0.006), has "glowing" features and a dark environment that make its look (0.168) and habitat (0.203) scores very high, giving it the highest total score (0.377), making it the most relevant monster overall. What an improvement! behavior look habitat Can we can replicate our results? Let's try another query and find out. query = { 'look': 'shapeshifting', 'habitat': 'varied landscapes', 'behavior': 'illusion creation' } query = { 'look': 'shapeshifting', 'habitat': 'varied landscapes', 'behavior': 'illusion creation' } id score look habitat behavior Mist Phantom 0.489574 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Zephyr Dancer 0.342075 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Whispering Shade 0.337434 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths id score look habitat behavior Mist Phantom 0.489574 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Zephyr Dancer 0.342075 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Whispering Shade 0.337434 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths id score look habitat behavior id id id score score score look look look habitat habitat habitat behavior behavior behavior Mist Phantom 0.489574 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Mist Phantom Mist Phantom 0.489574 0.489574 Ethereal, fog-like humanoid with shifting features Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Lures travelers astray with illusions and whispers Zephyr Dancer 0.342075 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Zephyr Dancer Zephyr Dancer 0.342075 0.342075 Graceful avian creature with iridescent feathers Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Creates mesmerizing aerial displays to attract mates Whispering Shade 0.337434 Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Whispering Shade Whispering Shade 0.337434 0.337434 Shadowy, amorphous being with glowing eyes Shadowy, amorphous being with glowing eyes Dark forests and abandoned buildings Dark forests and abandoned buildings Feeds on fear and whispers unsettling truths Feeds on fear and whispers unsettling truths Great! Our outcomes are excellent again. What if we want to find monsters that are similar to a specific monster from our dataset? Let's try it with a monster we haven't seen yet - Harmonic Coral. We could extract attributes for this monster and create query parameters manually. But Superlinked has a with_vector method we can use on the query object. Because each monster's id is its name, we can express our request as simply as: could with_vector app.query( monster_query.with_vector(monster, "Harmonic Coral"), **default_weights, limit=LIMIT ) app.query( monster_query.with_vector(monster, "Harmonic Coral"), **default_weights, limit=LIMIT ) id score look habitat behavior Harmonic Coral 1 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.402288 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Aqua Wraith 0.330869 Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents id score look habitat behavior Harmonic Coral 1 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.402288 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Aqua Wraith 0.330869 Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents id score look habitat behavior id id id score score score look look look habitat habitat habitat behavior behavior behavior Harmonic Coral 1 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Harmonic Coral Harmonic Coral 1 1 Branching, musical instrument-like structure with vibrating tendrils Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.402288 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Dreamweaver Octopus Dreamweaver Octopus 0.402288 0.402288 Cephalopod with tentacles that shimmer like auroras Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Influences the dreams of nearby creatures Aqua Wraith 0.330869 Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents Aqua Wraith Aqua Wraith 0.330869 0.330869 Translucent humanoid figure made of flowing water Translucent humanoid figure made of flowing water Rivers, lakes, and coastal areas Rivers, lakes, and coastal areas Shapeshifts to blend with water bodies and controls currents Shapeshifts to blend with water bodies and controls currents The top result is the most relevant one, Harmonic Coral itself, as expected. The other two monsters our search retrieves are Dreamweaver Octopus and Aqua Wraith. Both share important thematic ( attribute ) elements with Harmonic Coral: attribute Aquatic habitats (habitat) Ability to influence or manipulate their environment (behavior) Dynamic or fluid visual characteristics (look) Aquatic habitats ( habitat ) habitat Ability to influence or manipulate their environment ( behavior ) behavior Dynamic or fluid visual characteristics ( look ) look Attribute weighting Suppose, now, that we want to give more importance to the look attribute. The Superlinked framework lets us easily adjust weights at query time. For easy comparison, we'll search for monsters similar to Harmonic Coral, but with our weights adjusted to favor look . look look weights = { "look_weight": 1.0, "habitat_weight": 0, "behavior_weight": 0 } app.query( monster_query.with_vector(monster, "Harmonic Coral"), limit=LIMIT, **weights ) weights = { "look_weight": 1.0, "habitat_weight": 0, "behavior_weight": 0 } app.query( monster_query.with_vector(monster, "Harmonic Coral"), limit=LIMIT, **weights ) id score look habitat behavior Harmonic Coral 0.57735 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Thornvine Elemental 0.252593 Plant-like creature with a body of twisted vines and thorns Overgrown ruins and dense jungles Rapidly grows and controls surrounding plant life Plasma Serpent 0.243241 Snake-like creature made of crackling energy Electrical storms and power plants Feeds on electrical currents and can short-circuit technology id score look habitat behavior Harmonic Coral 0.57735 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Thornvine Elemental 0.252593 Plant-like creature with a body of twisted vines and thorns Overgrown ruins and dense jungles Rapidly grows and controls surrounding plant life Plasma Serpent 0.243241 Snake-like creature made of crackling energy Electrical storms and power plants Feeds on electrical currents and can short-circuit technology id score look habitat behavior id id id score score score look look look habitat habitat habitat behavior behavior behavior Harmonic Coral 0.57735 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Harmonic Coral Harmonic Coral 0.57735 0.57735 Branching, musical instrument-like structure with vibrating tendrils Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Creates complex melodies to communicate and influence emotions Thornvine Elemental 0.252593 Plant-like creature with a body of twisted vines and thorns Overgrown ruins and dense jungles Rapidly grows and controls surrounding plant life Thornvine Elemental Thornvine Elemental 0.252593 0.252593 Plant-like creature with a body of twisted vines and thorns Plant-like creature with a body of twisted vines and thorns Overgrown ruins and dense jungles Overgrown ruins and dense jungles Rapidly grows and controls surrounding plant life Rapidly grows and controls surrounding plant life Plasma Serpent 0.243241 Snake-like creature made of crackling energy Electrical storms and power plants Feeds on electrical currents and can short-circuit technology Plasma Serpent Plasma Serpent 0.243241 0.243241 Snake-like creature made of crackling energy Snake-like creature made of crackling energy Electrical storms and power plants Electrical storms and power plants Feeds on electrical currents and can short-circuit technology Feeds on electrical currents and can short-circuit technology Our results all (appropriately) have similar appearances - "Branching with vibrating tendrils", "Plant-like creature with a body of twisted vines and thorns", "Snake-like". Now, let's do another search, ignoring appearance, and looking instead for monsters that are similar in terms of habitat and behavior simultaneously: habitat behavior weights = { "look_weight": 0, "habitat_weight": 1.0, "behavior_weight": 1.0 } weights = { "look_weight": 0, "habitat_weight": 1.0, "behavior_weight": 1.0 } id score look habitat behavior Harmonic Coral 0.816497 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.357656 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Mist Phantom 0.288106 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers id score look habitat behavior Harmonic Coral 0.816497 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.357656 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Mist Phantom 0.288106 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers id score look habitat behavior id id id score score score look look look habitat habitat habitat behavior behavior behavior Harmonic Coral 0.816497 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Harmonic Coral Harmonic Coral 0.816497 0.816497 Branching, musical instrument-like structure with vibrating tendrils Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Creates complex melodies to communicate and influence emotions Dreamweaver Octopus 0.357656 Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Dreamweaver Octopus Dreamweaver Octopus 0.357656 0.357656 Cephalopod with tentacles that shimmer like auroras Cephalopod with tentacles that shimmer like auroras Deep ocean trenches and underwater caves Deep ocean trenches and underwater caves Influences the dreams of nearby creatures Influences the dreams of nearby creatures Mist Phantom 0.288106 Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Mist Phantom Mist Phantom 0.288106 0.288106 Ethereal, fog-like humanoid with shifting features Ethereal, fog-like humanoid with shifting features Swamps, moors, and foggy coastlines Swamps, moors, and foggy coastlines Lures travelers astray with illusions and whispers Lures travelers astray with illusions and whispers Again, the Superlinked approach produces great results. All three monsters live in watery environments and possess mind-controlling abilities. Finally, let's try another search, weighting all three attributes differently - to find monsters that in comparison to Harmonic Coral look somewhat similar, live in very different habitats, and possess very similar behavior: weights = { "look_weight": 0.5, "habitat_weight": -1.0, "behavior_weight": 1.0 } weights = { "look_weight": 0.5, "habitat_weight": -1.0, "behavior_weight": 1.0 } id score look habitat behavior Harmonic Coral 0.19245 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Luminoth 0.149196 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Zephyr Dancer 0.136456 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates id score look habitat behavior Harmonic Coral 0.19245 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Luminoth 0.149196 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Zephyr Dancer 0.136456 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates id score look habitat behavior id id id score score score look look look habitat habitat habitat behavior behavior behavior Harmonic Coral 0.19245 Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Harmonic Coral Harmonic Coral 0.19245 0.19245 Branching, musical instrument-like structure with vibrating tendrils Branching, musical instrument-like structure with vibrating tendrils Shallow seas and tidal pools Shallow seas and tidal pools Creates complex melodies to communicate and influence emotions Creates complex melodies to communicate and influence emotions Luminoth 0.149196 Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Luminoth Luminoth 0.149196 0.149196 Moth-like creature with glowing wings and antenna Moth-like creature with glowing wings and antenna Dense forests and jungles with bioluminescent flora Dense forests and jungles with bioluminescent flora Emits soothing light patterns to communicate and attract prey Emits soothing light patterns to communicate and attract prey Zephyr Dancer 0.136456 Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Zephyr Dancer Zephyr Dancer 0.136456 0.136456 Graceful avian creature with iridescent feathers Graceful avian creature with iridescent feathers High mountain peaks and wind-swept plains High mountain peaks and wind-swept plains Creates mesmerizing aerial displays to attract mates Creates mesmerizing aerial displays to attract mates Great results again! Our two other retrieved monsters — Luminoth and Zephyr Dancer — have behavior similar to Harmonic Coral and live in habitats different from Harmonic Coral's. They also look very different from Harmonic Coral. (While Harmonic Coral's tendrils and Luminoth's antenna are somewhat similar features, we only down-weighted the look_weight by 0.5, and the resemblance between the two monsters ends there.) look_weight Let's see how these monsters' overall scores break out in terms of individual attributes: id look habitat behavior total Harmonic Coral 0.19245 -0.3849 0.3849 0.19245 Luminoth 0.052457 -0.068144 0.164884 0.149196 Zephyr Dancer 0.050741 -0.079734 0.165449 0.136456 id look habitat behavior total Harmonic Coral 0.19245 -0.3849 0.3849 0.19245 Luminoth 0.052457 -0.068144 0.164884 0.149196 Zephyr Dancer 0.050741 -0.079734 0.165449 0.136456 id look habitat behavior total id id id look look look habitat habitat habitat behavior behavior behavior total total total Harmonic Coral 0.19245 -0.3849 0.3849 0.19245 Harmonic Coral Harmonic Coral 0.19245 0.19245 -0.3849 -0.3849 0.3849 0.3849 0.19245 0.19245 Luminoth 0.052457 -0.068144 0.164884 0.149196 Luminoth Luminoth 0.052457 0.052457 -0.068144 -0.068144 0.164884 0.164884 0.149196 0.149196 Zephyr Dancer 0.050741 -0.079734 0.165449 0.136456 Zephyr Dancer Zephyr Dancer 0.050741 0.050741 -0.079734 -0.079734 0.165449 0.165449 0.136456 0.136456 By negatively weighting habitat_weight (-1.0), we deliberately "push away" monsters with similar habitats and instead surface monsters whose environments are different from Harmonic Coral's - as seen in Luminoth's and Zephyr Dancer's negative habitat scores. Luminoth's and Zephyr Dancer's behavior scores are relatively high, indicating their behavioral similarity to Harmonic Coral. Their look scores are positive but lower, reflecting some but not extreme visual similarity to Harmonic Coral. habitat_weight habitat behavior look some In short, our strategy of downweighting habitat_weight to -1.0 and look_weight to 0.5 but keeping behavior_weight at 1.0 proves effective in surfacing monsters that share key behavioral characteristics with Harmonic Coral but have very different environments and look at least somewhat different. habitat_weight look_weight behavior_weight Conclusion Multi-attribute vector search is a significant advance in information retrieval, offering more accuracy, contextual understanding, and flexibility than basic semantic similarity search. Still, our naive approach (above) - storing and searching attribute vectors separately, then combining results - is limited in ability, subtlety, and efficiency when we need to retrieve objects with multiple simultaneous attributes. (Moreover, multiple kNN searches take more time than a single search with concatenated vectors.) then To handle scenarios like this, it's better to store all your attribute vectors in the same vector store and perform a single search , weighting your attributes at query time. The Superlinked approach is more accurate, efficient, and scalable than the naive approach for any application that requires fast, reliable, nuanced, multi-attribute vector retrieval - whether your use case is tackling real world data challenges in your e-commerce or recommendation system... or something entirely different, like battling monsters. a single search Contributors Andrey Pikunov, Author Mór Kapronczay, Editor Andrey Pikunov, Author Andrey Pikunov, Author Mór Kapronczay, Editor Mór Kapronczay, Editor Originally published here. Originally published here . here
