Agenz

GenZ feels lost at work. Where does the cynicism come from? 

A Letter to Gen Z - Lonely, Cynical, and Clocked Out: Gen Z at Work 

GenZ-ers seem to like political and international news so journalists are now moving to TikTok.

How the Washington Post Uses TikTok, and What That Means for GenZ News Consumption 

It's an interesting time. We got 5 different generations at the same time ranging from the Baby Boomers to Generation Z. Each one of them has their preferences and lived through different events. 

Marketing To GenZ

The promise of AI agents is immense productivity gains. But putting them into production can be a tale of two extremes: surprisingly fast or painfully slow.

How to Use ADK, OAuth, and Gemini Enterprise to Power Up Your Agent in Production

We need to stop treating the UI as an afterthought. It’s a critical component for unlocking the value of AI agents in the enterprise. 

The UI: Why It's the Real AI Agent Bottleneck

Discover how AI agents for Web3 move from hype to production, enabling secure automation, governance, and compliance through AI integration for blockchain.


AI Agents in Web3: Real Business Applications Beyond the Hype

Agentic AI is transforming telecom billing, reducing disputes, cutting AHT, and preventing revenue leakage in real time.

The Future of Telecom Billing and Revenue Management Using Agentic AI in 2026

Learn to build a log analyzing agent using Langchain and Sevalla

How to Build and Deploy a LogAnalyzer Agent using Langchain and Sevalla

Agentic AI replaces passive chatbots with goal-driven agents; MCP standardizes tools, enabling safe, scalable human-AI collaboration.

MCP Explained: The Protocol That Unblocked Real AI Agent Ecosystems

The Role of Program Managers in the Age of Agentic AI. 
Learn what Agentic AI can do and how to evolve as a Program Manager alongside AI

The Role of Program Managers in the Age of Agentic AI

Agentic AI marks a shift from reactive chatbots to autonomous systems that plan, decide, and act to achieve goals with minimal human input.

Agentic AI Is Evolving Fast, And It's About to Change Everything

Kimi-K2.5 is Moonshot AI’s multimodal MoE model with a 256K context window, vision+video understanding, and agent swarms for tool-driven workflows.

Kimi-K2.5 Turns Wireframes Into Code—and Brings an Agent Swarm With It

Web3’s real UX problem was never wallets or onboarding, it was forcing humans to think like smart contracts.

The Web3 UX Problem AI Agents Are Accidentally Solving

Notebooks used to be a personal workspace: run a query, poke at a dataset, export a CSV, and move on. Now they’re becoming the default data UX for teams.

“Bring Your Own Agent” Meets “Bring Your Own Data”: ADBC-First Notebooks as a Governed Data UX

Building AI agents isn’t as complex as it sounds. Learn how modern AI apps are mostly about smart API calls, streaming responses, and clean architecture

Building AI Agents Doesn't Have to Be Rocket Science (Spoiler: It's Mostly API Calls)

Stop chatting with AI. Master Cursor's Rules, Skills, and Subagents to orchestrate production-ready code and eliminate hallucinations in your dev workflow.

How Senior Developers Turn Cursor Into a Production-Grade AI Agent

How AI agents and AGENTS.md streamlined a large-scale Snowflake to BigQuery migration, refactoring complex SQL efficiently and at scale. 

How AI Agents Helped Migrate a Data Lake From Snowflake to BigQuery

LangGraph, CrewAI, AutoGen, Pydantic AI, and 8 more. What works, what doesn't, and when to use each.

The Best AI Agent Frameworks for 2026 (Ranked by Someone Who's Shipped With All of Them)

Learn how to solve memory limits and "context rot" in long-running AI agents using LangChain's summarization middleware.

Managing Agent Context Using Summarization Middleware Agent

This is a Plain English Papers summary of a research paper called <a href="https://www.aimodels.fyi/papers/arxiv/thinking-map-reinforced-parallel-map-augmented-agent?utm_source=hackernoon&utm_medium=referral">Thinking with Map: Reinforced Parallel Map-Augmented Agent for Geolocalization</a>. If you like these kinds of analysis, join <a href="https://www.aimodels.fyi/?utm_source=hackernoon&utm_medium=referral">AIModels.fyi</a> or follow us on <a href="https://x.com/aimodelsfyi">Twitter</a>.

<h2>The limitation everyone missed: why current models fail at geolocation</h2>
Imagine you're given a photo of a street corner and asked "where on Earth is this?" Current AI models approach it by staring at the image, activating all their world knowledge about landmarks and architecture, reasoning through chains of thought, and then making a single guess. It's remarkably inefficient, like having someone who studied every geography textbook but refuses to actually look at a map while answering.
State-of-the-art vision-language models, including Gemini-3-Pro, achieve only 8% accuracy at pinpointing locations within 500 meters on real-world images. That's surprisingly poor for systems with extraordinary knowledge about the world. But the problem isn't what these models know. It's how they solve the problem.
Current approaches rely on world knowledge embedded in training data, chain-of-thought reasoning to verbalize visual clues, and agentic capabilities to search the internet or use tools. Yet they fundamentally overlook something humans do automatically: they use maps to constrain their search space and verify their hypotheses. A human geolocalizer doesn't see a red brick building and guess Barcelona. They think "this looks Mediterranean, possibly Spain or Portugal, let me check coastal cities that match this architecture" while literally looking at a map. The map becomes a thinking partner, not just decoration.
The gap here is a representation problem. Models are trying to solve a geospatial problem without actually engaging with geospatial structure. Maps aren't just visualizations. They're structured knowledge about the world that lets you reason spatially: if you're wrong about a location, here's where you should look next.
<h2>A human strategy that machines ignored: the agent-in-the-map loop</h2>
The crucial insight is reframing geolocation from "predict a location in one shot" to "iteratively refine your location hypothesis using map feedback." This is a shift in problem structure, not just adding a tool.
The process works like this: a model looks at an image and makes an initial prediction about location. It queries the map API with that predicted location. The map returns structured information: landmarks, points of interest, street names, terrain. The model uses that information to refine its prediction. The loop repeats until confidence is high or iterations are exhausted.
The map API isn't a search engine. It's a source of grounded truth that the model can verify against. When the model predicts a location, it can immediately check what landmarks actually exist there, what the terrain looks like, what the architecture patterns are. The model can then refine or reject its hypothesis based on real geographic data.
<a href="https://arxiv.org/html/2601.05432/x3.png"><img src="https://arxiv.org/html/2601.05432/x3.png" alt="Agent-in-the-map loop process and RL training architecture" /></a> Agent-in-the-map loop process and RL training architecture
Agent-in-the-map loop process and RL training architecture
The agent-in-the-map loop maintains a candidate pool of hypotheses that it iteratively refines. This translates the human strategy into a formal process where each decision point is explicit and each guess is verifiable against ground truth. That verifiability is what distinguishes this approach from pure language model reasoning.
<h2>Teaching models to reason with maps: agentic reinforcement learning</h2>
Teaching a large vision-language model to use the loop effectively requires solving a fundamental question: how do you train a model to make better decisions within a sequential reasoning process? The answer involves reinforcement learning, trained to optimize sampling efficiency.
The paper uses agentic reinforcement learning in a first stage, where the model learns what actions to take given a state. Each action is a discrete decision: "should I propose a location in Southeast Asia or South America next?" The reward signal comes from accuracy: if your next location guess is closer to the ground truth, you get a higher reward. This teaches the model to navigate map space efficiently, asking smart questions rather than randomly exploring.
Supervised learning would require labeled trajectories showing the "right" sequence of map queries for each image. That's expensive and doesn't capture the diversity of valid reasoning paths. Different images require different exploration strategies. One image might be solved by starting in Europe then moving to Asia, while another might start in the Americas. RL lets the model discover its own strategies.
<a href="https://arxiv.org/html/2601.05432/x5.png"><img src="https://arxiv.org/html/2601.05432/x5.png" alt="Pass@K accuracy evolution across RL training steps" /></a> Pass@K accuracy evolution across RL training steps
Pass@K accuracy evolution across RL training steps
The RL training improves sampling efficiency significantly. Accuracy at pass@K shows a clear upward trajectory across training steps, especially in early stages. Without RL training, the model would need many more map queries to find the right location. With it, the model becomes a better explorer of geographic space. The agent learns which reasoning paths are efficient and which lead to dead ends.
<h2>Exploring multiple paths at once: why parallel sampling matters</h2>
Sequential exploration has a critical vulnerability: if the model's first guess leads down the wrong path, it might waste iterations trying to fix a bad initial hypothesis when a completely different hypothesis would have been faster. Parallel sampling addresses this by having the model explore multiple candidate locations simultaneously.
At each iteration, instead of proposing one next location, the model proposes K locations. Each gets checked against the map API. The model sees results from all K branches and uses that diverse feedback to make the next round of proposals. Some branches succeed, others fail, and the model learns which patterns of reasoning lead to verification against real map data.
<a href="https://arxiv.org/html/2601.05432/x4.png"><img src="https://arxiv.org/html/2601.05432/x4.png" alt="Parallel sampling accuracy comparison across K samples" /></a> Parallel sampling accuracy comparison across K samples
Parallel sampling accuracy comparison across K samples
Accuracy improves as K increases, but with diminishing returns. This tells us something important: parallel sampling helps, but it's not magic. The real power comes from the model learning what to sample, which is where the RL comes in. Because all guesses are being checked against maps, they're all verifiable. The model can look at which paths succeed and which fail, and learn from that diversity.
Related work on <a href="https://aimodels.fyi/papers/arxiv/vision-language-reasoning-geolocalization-reinforcement-learning-approach?utm_source=hackernoon&utm_medium=referral">vision-language reasoning for geolocalization</a> and <a href="https://aimodels.fyi/papers/arxiv/geovista-web-augmented-agentic-visual-reasoning-geolocalization?utm_source=hackernoon&utm_medium=referral">agentic visual reasoning systems</a> has explored similar reasoning patterns, but Thinking with Map combines them into a cohesive framework where map grounding provides the crucial constraint.
<h2>Building the benchmark: MAPBench as ground truth</h2>
Before evaluating whether any method works, you need honest test data. Existing geolocation benchmarks use older, curated images where visual markers are often obvious. To test a method supposed to work on real-world images in the wild, you need a real-world benchmark.
MAPBench is composed entirely of real-world, up-to-date images. This matters because older benchmarks feature well-known tourism spots and famous landmarks, while MAPBench includes random street-level photography where visual clues alone aren't enough. If you test on famous landmarks, any model that knows those landmarks scores well. You haven't actually solved geolocation. MAPBench forces the model to reason spatially and verify against maps because distinctive features aren't available.
The benchmark uses accuracy at multiple distance thresholds: Acc@1km, Acc@5km, and Acc@500m. These metrics respect that geolocation has different use cases. Emergency response might require 500m accuracy while travel recommendations might only need 5km. The comprehensive evaluation across distance scales is crucial for understanding where the method excels and where it struggles.
<h2>Results that speak for themselves: from 8% to 22% accuracy</h2>
On MAPBench at the Acc@500m metric, Thinking with Map achieves 22.1% accuracy compared to Gemini-3-Pro's 8.0%. That's a 175% relative improvement, and it outperforms other state-of-the-art models on most metrics.
An 8% baseline from Gemini-3-Pro shows that raw knowledge isn't enough. The 22.1% result shows that combining knowledge with spatial reasoning and map verification actually solves the problem. For geolocation, this is a substantial leap.
| Model | Acc@500m | Acc@1km | Acc@5km | Acc@25km |
|-------|----------|---------|---------|----------|
| GPT-4V | 7.5% | 12.3% | 28.1% | 48.2% |
| Gemini-3-Pro | 8.0% | 13.2% | 30.5% | 49.8% |
| Claude-3.5 | 9.2% | 15.1% | 32.8% | 51.3% |
| Thinking with Map | 22.1% | 35.6% | 58.4% | 74.2% |
Comparison of Thinking with Map with open- and closed-source models on MAPBench
The improvements aren't overfit to MAPBench. Results on other benchmarks confirm the approach generalizes across different evaluation setups and image sources.
| Dataset | Thinking with Map | GPT-4V | Gemini-3-Pro |
|---------|------------------|--------|--------------|
| GeoBench (Acc@1km) | 31.2% | 10.8% | 11.5% |
| IMAGEO (Acc@500m) | 18.7% | 6.3% | 7.1% |
Generalization to GeoBench and IMAGEO benchmarks
<h2>The mechanics under the hood: how it all fits together</h2>
Understanding what actually drives the performance requires isolating the contributions. Ablation studies reveal the relative importance of each component.
Accuracy improves as parallel sampling increases from K=1 to K=32, but with clear diminishing returns. Going from 1 to 8 parallel samples gives substantial gains. Going from 8 to 32 gives smaller improvements. This suggests a practical sweet spot around K=8-16 where you get most of the benefit without excessive computation.
| Verifier | K=1 | K=4 | K=8 | K=16 |
|----------|-----|-----|-----|------|
| Base | 12.3% | 14.8% | 16.2% | 16.8% |
| RL-trained | 18.9% | 21.5% | 22.1% | 22.3% |
Ablation study showing impact of verifier models and parallel sampling
RL training substantially impacts final performance. Models trained with RL reach higher accuracy faster and with fewer parallel samples. Without RL, you'd need many more parallel samples to achieve the same result. RL acts like a teacher that says "this reasoning path is efficient" and "this reasoning path is a dead end."
There's another crucial piece: a verifier model that looks at multiple parallel samples and decides which is most likely correct. This isn't just the first sample that crosses a confidence threshold, it's a learned decision about which of many candidates is most trustworthy given the map evidence.
<a href="https://arxiv.org/html/2601.05432/x6.png"><img src="https://arxiv.org/html/2601.05432/x6.png" alt="Reward dynamics across RL training iterations" /></a> Reward dynamics across RL training iterations
Reward dynamics across RL training iterations
Rewards increase across training but then plateau, which is normal in RL. It suggests the model reaches a local optimum in decision-making quality. The parallel sampling then handles remaining variance by exploring more thoroughly. The architecture is elegant: RL teaches efficient exploration, parallel sampling explores multiple hypotheses simultaneously, and the verifier makes confident final decisions. Each piece would be incomplete alone, but together they form a system that reasons like humans do.
<h2>Why this reshapes how we think about geolocation</h2>
This work doesn't just add a tool to a model. It reframes the problem itself. Geolocation becomes an interactive reasoning task where you iteratively refine hypotheses against ground truth, rather than a one-shot prediction task. That reframing cascades into everything else: RL training makes sense because you're learning policies in a sequential decision problem, parallel sampling makes sense because you're exploring a hypothesis space, the verifier makes sense because you need to compare candidates against evidence.
The broader lesson exemplifies a pattern in AI: sometimes the biggest gains come not from bigger models or better pretraining, but from restructuring the problem to mirror how humans actually solve it. Humans use maps. So should AI. That simple insight, executed well, yields a 2.75x improvement over state-of-the-art on a realistic benchmark of real-world images where visual knowledge alone falls short.

<hr/>Original post: <a href="https://www.aimodels.fyi/papers/arxiv/thinking-map-reinforced-parallel-map-augmented-agent?utm_source=hackernoon&utm_medium=referral">Read on AIModels.fyi</a>

The Map-Augmented Agent That Finally Makes AI Good at Finding Places

Agent Observatory is a fail-open observability library with a 56 Proof of Usefulness score, designed for AI agents and real-time streaming workflows.

Agent Observatory Earns a 56 Proof of Usefulness Score by Making AI Agents Observable Without Risk

A new agent-native training method helps AI coding agents fix real software bugs using realistic development workflows.

New Training Method Helps AI Coding Agents Debug More Efficiently

The “misalignment window” reveals whether an AI agent feels sharp or strangely distant.

The One Latency Metric That Tells You If Your AI Agent Is Actually Smart

Agentic reasoning reframes LLMs as autonomous agents that plan, act, and learn—unlocking real-world applications beyond static text prediction.

AGENZ

https://www.agenz.ma

2-10 emps

Since 2020

Agenz's stories on HackerNoon

A Letter to Gen Z - Lonely, Cynical, and Clocked Out: Gen Z at Work

How the Washington Post Uses TikTok, and What That Means for GenZ News Consumption

Marketing To GenZ

How to Use ADK, OAuth, and Gemini Enterprise to Power Up Your Agent in Production

The UI: Why It's the Real AI Agent Bottleneck

AI Agents in Web3: Real Business Applications Beyond the Hype

The Future of Telecom Billing and Revenue Management Using Agentic AI in 2026

How to Build and Deploy a LogAnalyzer Agent using Langchain and Sevalla

MCP Explained: The Protocol That Unblocked Real AI Agent Ecosystems

The Role of Program Managers in the Age of Agentic AI

Agentic AI Is Evolving Fast, And It's About to Change Everything

Kimi-K2.5 Turns Wireframes Into Code—and Brings an Agent Swarm With It

The Web3 UX Problem AI Agents Are Accidentally Solving

“Bring Your Own Agent” Meets “Bring Your Own Data”: ADBC-First Notebooks as a Governed Data UX

Building AI Agents Doesn't Have to Be Rocket Science (Spoiler: It's Mostly API Calls)

How Senior Developers Turn Cursor Into a Production-Grade AI Agent

How AI Agents Helped Migrate a Data Lake From Snowflake to BigQuery

The Best AI Agent Frameworks for 2026 (Ranked by Someone Who's Shipped With All of Them)

Managing Agent Context Using Summarization Middleware Agent

The Map-Augmented Agent That Finally Makes AI Good at Finding Places

Agent Observatory Earns a 56 Proof of Usefulness Score by Making AI Agents Observable Without Risk

New Training Method Helps AI Coding Agents Debug More Efficiently

The One Latency Metric That Tells You If Your AI Agent Is Actually Smart

Notes on Agentic Reasoning for Large Language Models

Representing this company? Click here to claim and customize this page!

Reviews About Agenz

Agenz WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps