Woedenis in the Industrial Robotics Industry

Woeden

Scientific American, Volume XXXVI., No. 8, February 24, 1877 by Various, is part of the HackerNoon Books Series. You can jump to any chapter in this book . The Wetli Mountain Railroad and Its Disastrous Trial Trip here THE WETLI MOUNTAIN RAILROAD AND ITS DISASTROUS TRIAL TRIP. Among the various means proposed of late years for building lines of railroad on the steep slopes of mountains, that of M. Wetli, of Zurich, Switzerland, has attracted considerable attention from European engineers. We have already laid before our readers the system of central toothed rails used on the Righi and other mountain roads in Europe. In the Wetli system, instead of this rail and the pinion on the vehicle engaging it, there is a drum having a helicoidal thread which engages with triangular rails. This drum is attached to the locomotive. The construction will be readily understood from the illustrations given herewith, which we take from . The thread on the drum is precisely that which would be formed could a rail similar to one of the central angular rails be wrapped around it; so that it always is in contact with the mid rails, and necessarily prevents any bodily sliding or rolling of the vehicles over the regular track when the drum is held motionless. The V-shaped mid rails are securely fastened to horizontal iron ties, which rest on wooden traverses. The angle of the V is 50°; the distance between any two traverses is 2.8 feet. La Nature The locomotive has three coupled axles, on the mid one of which the drum is attached so as to be raised or lowered to engage the rails at the will of the engineer: it being possible to cause it to act on the rails with a pressure of 3.7 tons. The diameter of the drum is 2.14 feet. Its spiral thread is of steel, very solidly attached, and so made as to grip the rails to a distance of 0.6 inch below the level of the track. In order to insure this contact, on the drum axle are two pulleys which run on the exterior road, and of which the diameter determines the depth of the hold of the threads. These pulleys are 34.7 inches in diameter, while the driving wheels are very slightly in excess, to provide for the use of tyres. M. Wetli's invention, as we have described it, was placed between Woedensweil and Einsiedlen, Switzerland. The difference in altitude between these points is 1,513 feet, the distance 9.6 miles. The grade is from 4 to 5 per cent over the first six miles of the way, and subsequently decreases to 1 per cent. The Wetli railroad was established last October only on the heavy grade, that is, the first six miles. Early in November, trial trips were made which did not prove satisfactory. Sometimes the drum thread gripped the triangular rails properly and acted well; again it would wedge itself upon them, and often would simply roll over their tops without engaging at all. After the first trials, during which very many of the rails were broken, M. Welti re-adjusted the drum thread. Finally, he concluded that he had overcome all difficulties in his apparatus; and accordingly a formal trial was arranged on November 30. For about four and a half miles of the ascent the drum worked well; and the hoarfrost, with which the rails were thickly covered, showed good contact. Afterward it worked irregularly; but the station of Schindelleghi, a distance of five miles, was reached without accident, the locomotive dragging a car loaded with 20 tons of rails. It was then attempted to make the descent by the aid of the helicoidal drum; but this jumped the rails, and broke them almost immediately. By the aid of back pressure of steam and brakes, the locomotive was stopped. Then, unfortunately, the engine was started again; but hardly had the descent been resumed when it was evident that the drum was not holding, and that the speed was accelerating rapidly. Brakes and steam were both found useless, and the engine went tearing over the rails at the rate of a mile a minute. Of the fourteen persons in the vehicles, three were thrown out and killed, and the rest were more or less seriously injured. The heavily loaded car left the track, and tore up both central and side rails until its coupling broke. The engineer, with great intrepidity, clung to his engine, coolly giving signals to open switches so that the locomotive might run upon the level track and so expend its momentum; but the engine left the rails at a sharp curve, destroyed the track for about a hundred feet, and finally stopped a mass of ruins, with its brave engineer mortally wounded. Whether the Wetli system can be made to work as intended by the inventor is regarded as doubtful by the engineers who have examined into the causes of the disaster. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Various (2006). Scientific American, Volume XXXVI., No. 8, February 24, 1877. Urbana, Illinois: Project Gutenberg. Retrieved https://www.gutenberg.org/cache/epub/19406/pg19406-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at www.gutenberg.org https://www.gutenberg.org/policy/license.html

THE WETLI MOUNTAIN RAILROAD AND ITS DISASTROUS TRIAL TRIP

Things To Make by Archibald Williams is part of the HackerNoon Books Series. You can jump to any chapter in this book . A WOODEN WORKBOX here XXXIII. A WOODEN WORKBOX. The box illustrated by Fig. 181 was copied from an article of Norwegian manufacture. Its construction is an extremely simple matter, provided that one can get a piece of easily bent wood (birch, for instance), not exceeding 3/16 inch in thickness, for the sides. [Illustration: FIG. 180.—Showing how to draw an ellipse.] [Illustration: FIG. 181.—Norwegian workbox.] The bottom of the box is made of 5/16 or 3/8 inch wood, cut to an oval or elliptical shape. To mark out an ellipse about 8 inches long and 5-1/2 inches wide—this will be a. convenient size—stick two pins into the board 5-1/8 inches apart, pass a loop of thread 14 inches in circumference round these, and run the point of a pencil round the pins in the path which it has to take when confined by the slack of the loop (Fig. 180). Fret-saw along the line. The wood strip for the side is 4-1/2 inches deep, and 1-1/2 inches longer than the circumference of the bottom. The ends are thinned off somewhat, as shown in Fig. 181, to prevent the lap having a clumsy appearance, and the surface is smoothed all over with sandpaper. Bore a number of small nail holes 3/16 inch from one edge, and then steam the wood over a big saucepan or other suitable vessel until it is quite lissom. When attaching the side piece to the bottom, begin at the middle, and work first towards what will be the inside end of the lap, and then towards the outside end. Nails are driven in through the holes already drilled. When nailing is finished, clip the top of the overlap with a hand-vice or screw spanner, to prevent the tops of the ends sliding over one another, and bore a line of holes l/4 inch apart, and at the same distance from the outer end. Fine copper wire drawn to and fro through alternate holes from one end of the row to the other and back again, will secure the joint. The lid overlaps the side 1/4 inch in all directions and has a square notch cut in it at one end to pass under the piece A, and at the other a deeper, circular-ended nick to enable it to pass over the key B when that is turned into the position shown in the illustration. A is cut out of 1/4-inch wood; B, in one piece, out of 1/2-inch. Their length under the heads exceeds the inside depth of the box by the thickness of the lid. A is affixed rigidly to the side by small screws or wire, while B must be attached in a manner, which will allow the head to rotate. Cut two nicks round the shank, and two horizontal slots at the same height through the end of the box. A couple of brass rings must then be procured of such a size that, when flattened into a somewhat oval shape, they will project beyond the slots sufficiently to allow a piece of wire to pass through them and prevent their being drawn back again. Quarter-inch wood will do for the lid. A handle is made out of a couple of inches of small cane bent into a semicircle, let through the lid at each end, glued, and cut off flush. The exterior may be decorated by a design in poker-work, or be stained and varnished. This is left to the maker’s discretion. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Archibald Williams (2005). Things To Make. Urbana, Illinois: Project Gutenberg. Retrieved https://www.gutenberg.org/cache/epub/14664/pg14664-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

A WOODEN WORKBOX

In this interview, I speak to author and engineer on the potential implication Artificial Intelligence could possibly have on Fiction Writing. We speak about his new , AI-written books, and how can potentially alter the creative process of every writer. Michael Woudenberg book AI Edem: Hey Michael, it’s incredible to have you! Could you share more about yourself, your professional qualifications, and such Hi, it’s wonderful to be part of the Tech Writer’s Stack, and I’ve loved the incredible network and collaboration that Substack is allowing us to build, which I found largely absent from other blog spaces.  My background is a menagerie of experiences across multiple domains and disciplines.  I call myself an aspiring ; a jack of all trades and master of none, but more often better than a master of one. Michael: Polymath is in systems engineering, but it’s more like systems thinking.  I like to say that is more than just engineering and consists of three things: My core background Systems Thinking First, insatiable curiosity. Second, the humility to accept we just don’t know that much. Third is intentional reframing of a problem or idea to see if we really do understand it. I’ve got a Master's Degree in Systems Engineering from Johns Hopkins and a diverse background in advanced technologies such as autonomy, artificial intelligence, blockchain, cyber, aerospace, national security, and weapon systems across a variety of companies from tech startups to Fortune 150 companies. I also recently founded , which provides Systems Innovation to unlock the full potential across multiple domains and disciplines.  As mentioned earlier, I am also the author on Substack of , which provides counterintuitive insights from technology, innovation, philosophy, psychology, and more. Polymathic Disciplines Polymathic Being Edem: You recently announced your new book titled Paradox, Could tell us about it? that explores AI and dives into what it means to be human.  The tagline is: In the battle over Advanced AI, will we lose our humanity or learn what truly makes us human? Michael: Paradox is a science fiction novel It really explores the paradox of being human and what it would take to actually achieve a sentient artificial intelligence, along with all the challenges and potential consequences. My goal is that I want people to wonder which side they’d take in the end because I’m still not sure myself. Here’s a fun video trailer that gives a nice summary. https://youtu.be/k4_Ej2B2ZV4?si=jPnTuTjpVaW490-Q&embedable=true Edem: For the sake of Completion, could you define what fiction means? Fiction is the art of narrative that is not purely grounded in fact. , though, is where science fiction meets science fact. The book idea started about five years ago when I was trying to figure out how to take the psychology, sociology, and theology of being human and make it more interesting to read. Michael: Paradox AI makes a great plot device to work with because the goal, or the worry, is if we achieve artificial general intelligence (AGI), humanity as we know it is in trouble. AGI simply means that AI is as good or better than humans.  But to do that, you need to learn a lot about what it means to be human, and we are a lot more complex than we fully understand.  AI can only handle the logical right now, and humans are built on top of an emotional brain, which is also connected to a hormonal or gut brain.  Let’s just say there’s a lot to explore there. Edem: I’m sure you get asked this a lot, but did you make use of AI to write it? I did use AI to help write it.  But then again, any author who is using a computer with a word processor and a search engine is using AI to write.  What is slightly different is that ChatGPT 4.0 helped me to kick it off.  I had already crafted the entire outline, the ideas, and everything.  I just asked GPT to put together a narrative on a few topics. Michael: But the truth is that AI is a terrible author! It takes so much time and effort to craft the prompts right for any results.  Anything it created, I ripped apart and rewrote completely.  But, it helped me get off the blank page problem that so many authors are blocked by. It was a terrible writer, but it got started. me I used AI with Grammarly and WordTune, as well as Bing with GPT, to help me do editing and research. It was useful for coming up with names for characters and organizations.  For example, Hyperion Defense, one of the key organizations defending humanity, was recommended by ChatGPT due to the connection to the Greek myth of the Titan Hyperion, known for his watchfulness and protection. Edem: Let’s talk about the potential of AI in fiction. How do you envision AI could change the fiction writing landscape? Right now, AI is a terrible author.  It’s okay with simple descriptions, research, character definition, etc. It really can’t do longer form narrative well. It’s something that is explored in , which is titled Almost Intelligent.  What the characters find is that AI is good at mimicry; it just isn’t very good at creativity. Michael: Paradox in Chapter 3 A more painful conversation is that those who are worried about AI replacing them probably have a point.  The key to not being replaceable is to have a skillset or, more specifically, an understanding of what makes humans unique that can’t be algorithmized away.\ I do worry that a lot of weak writers will to crank out bad content.  We saw this recently with a book someone published literally two days after the Maui fire about how the fire was indicative of climate change.  When you got into the text, it was not just bad; it was horrible. I’m not even sure the fellow took the time to read it before he pushed it to Amazon. start using the LLMs Edem: If you were to rewrite your book, how would you use Text models to facilitate your creative process? Well, I wouldn’t rewrite this book, but I am about to kick off book two.  It’s titled Integration and explores the AI arc after the calamity of . It’s going to be exploring a lot of concepts more broadly, like race, gender, and nationality, among others, but back to your question, I don’t think I’m going to change much. Michael: Paradox I’ll use AIs like ChatGPT, WordTune, and Bing to help me with my research, scene descriptions, and similar, but these tools are just not at a place where they can write to the level I need them to. Edem: Is it possible for Large Language Models to ever be capable of creating Literary masterpieces like Harry Potter and Irobot? LLMs will be able to mimic those masterpieces.  But it will clearly be a mimic. I can’t see them being able to create something new and unique. Not without pretty heavy human involvement. Michael: Another thing to consider is that, right now, AI can’t learn. It can be trained, but once the training is done, it’s static again. Humans never stop learning though and social learning is one of our superpowers where we build off of one another. Humans learn and create through social connections. Moreso, what I found in the book that limited AI was that it was purely logical and could only fake emotion. Humans are smart because we are emotional, not so much that we are logical. That’s the piece that had to fall into place in to actually achieve AGI, but that’s just not something we are anywhere close to. So, no, AI is nowhere near capable of creating masterpieces. At least not to the trained eye Paradox Edem: Let’s take a look at ethics. If hypothetically, a book is co-authored with the aid of AI, who should take credit for the work and why? I’ll ask the question back a different way. If I went to a library and a librarian helped me do research, edited some of my ideas, etc., how much credit should they get?  The thing with AI like ChatGPT is that they are only responding with mathematically structured algorithms returning statistically predictive textual completion. Michael: They aren’t creating per se.  They certainly aren’t being creative. I will acknowledge AI was a useful tool, among many others, but there’s a ton of human input to the book, which just goes back to the idea from earlier on how we are social learners. I can honestly say there is nothing in my book that isn’t first mine and then aided by people and tools. Edem: What are your thoughts on a future where AI systems can efficiently imitate the literary technique of popular authors and potentially extend their literary legacy? I mean, we are doing that in some ways already.  They did that with Carrie Fisher in the 9th . But I think even if they could imitate, would we, as humans, appreciate it as much? Michael: Star Wars film I think of J.R.R. Tolkien. He has hundreds of manuscripts and world-building concepts he never wrote into books.  That’s why the Silmarillion is so clunky. It was never formed and polished.  Could AI help us with a better version of that? But do we want a better version of that? Or is his true legacy that he nearly wholesale invented the fantasy literary world as we know it?  Thousands of new authors are imitating the worlds he created, using his concepts of elves, dwarves, knights, and wizards to this day.  His legacy is seen in all the authors who were moved to build on it.  AI can only copy; it could never do what those human authors did with Tolkien’s legacy. Edem: Collaboration with AI seems inevitable. How best can writers incorporate AI into their creative process without outsourcing their potential creativity? I’d say that if you outsource your creativity, you’re not really creative. But AI can be incredibly helpful, as I mentioned earlier. You can even use it to provide counter-arguments or recommend new phrasing.  In drafting my second book, I’ve been asking ChatGPT to recommend changes to dialog to make my antagonist more antagonizing in a sneering, condescending way.  That helps get my brain juices flowing and I then craft it in my own way. Michael: But this isn’t any different in some ways from bringing up a personal experience and trying to express it.  AI just helps the process flow a bit faster. It still requires me to feed very specific prompts to get focused answers that are usable.  Without my creative input, its output is rubbish. It’s good as an admin, an editor, and an idea generator in the writing process. In publishing AI, behind distributors like Amazon’s Kindle Direct Publishing, it actually gives me the ability to write, design, publish, and fulfill orders in ways that didn’t exist 15 years ago. Edem: Looking \Michael: forward to the distant future. What parts of the entire writing process (writing, publishing, editing, marketing, and publishing) do you think AI tools could be most useful? AI in marketing through places like Meta’s ad space makes writing ads and finding eyes and clicks so easy it’s incredible. AI has literally democratized publishing to where anybody can do it without the archaic and nepotistic publishers and without paying many thousands of dollars.  All in, thus far, I’ve probably invested $2,000 to write and publish my book, including some marketing and more.  Twenty years ago, that wouldn’t have been possible. Edem: What advice would you give to a budding writer looking to write and publish their work in the current state of AI? Don’t let AI replace your writing. You have to be creative, unique, and thought-provoking.  If you’re writing isn’t any better than an AI, I’d look hard at improving your craft and finding out where you can be unique. AI also offers writers a unique opportunity to step out and above the average by understanding and leveraging the unique capabilities that humans can leverage over machine AI. Michael: AI is an enabler, not a replacement, for good writing. Those who use it to do the work for them will never compete with those who use it to make their work better. Edem: Well, that’s a wrap, Michael! Do you have any closing words? A lot of people are worried about AI.  My biggest worry isn’t AI but how we react to AI. That’s the thread I pulled in Paradox that unleashed the most destruction.  I didn’t need to have AI kill people; I just had the AI in the story unlock the fears and ignorance so that it triggered a human societal auto-immune response, and the humans did all the dirty work. Michael: But that’s not how it has to be, and that’s the true underlying lesson in the book.  Because it really does come down to that tagline: In the battle over Advanced AI, will we lose our humanity or learn what truly makes us human? If you are interested in more of these topics, please subscribe at PolymathicBeing.com You can find the book on . You can also watch a short video (2-½ minutes) on AI as a co-author below on YouTube, which you can share with others. Amazon

How AI Is Impacting Fiction Writing: An Interview with Michael Woudenberg 

After completing one project, we’re often in a hurry to start a new one. But why the rush? If you complete a quality project, recycle it: iterate on it and create your own mini-projects along the way. I do this myself. The process is helping to build my skills and solidify concepts—and it can help you, too. So in this article, I’ll share my process for creating programming projects. Let’s start with an approach I’ve adopted from the famous basketball coach, John Wooden. It’s called the “whole-part method.”[1] 1. The Whole-Part Method Coach Wooden used his “whole-part method” when teaching his team a particular basketball play. First, he’d show the team the entire play from start to finish. Then, he’d break the play down and the team would work on each part.[1] It’s a simple—but extraordinarily successful—learning strategy that you can apply to programming. It’ll help you nail the core competencies of a concept or technology. It'll also help you to see how these individual parts fit into a broader context. Start with a project that you’ve already completed, much like Coach Wooden showed his team a new basketball play from start to finish. Say you just completed a React project, where a user enters todos. The todos are displayed on your screen. Let’s use this as an example of a base project. Here’s how it works. Focus on Each Part Next, break the project down into smaller projects, just as Coach Wooden’s team would break down and drill each part of a new play. These are several mini-projects you could create from this React “todo” project: Implement the project using functions. Implement the project using classes. Add CSS to one version of the project. Add functionality to remove a todo. Notice that each mini-project centers around a core concept of React, like using classes and functions. This is one way to identify mini-projects: . Alternatively, you can create a project around something specific that you want to get better at. identify core concepts and design a small project around them Drill It and Internalize It Note that you’re not re-completing the entire project each time. Instead, you’re . The aim of this practice is to . focusing on one thing take something that has multiple parts, and break it down and drill each part Recall that one mini-project focuses on classes: write some classes in order to get a few user-entered “todos” to show up. Don’t add CSS. Don’t add the functionality to remove a todo. It’s pretty bare-bones, but that’s the point: you’re concentrating solely on writing classes. Likewise, when you work on the CSS mini-project focus exclusively on CSS. Take one version of the project that you’ve already completed and add CSS to it. Isolating specific aspects of a project and drilling them is a great way to really understand and internalize core concepts and build skills. It’ll also point out areas of weakness. If you struggle with classes, then you’ve identified an area to work on. Once you’ve completed a mini-project—and this is the critical point—take time to reflect: What did you do well? Did you try something different? What’s something that could’ve gone better? Are there any stumbling blocks? The idea is to look at your work objectively and analyze it: give yourself . If you struggled with a project, then let a few days pass and repeat it. This is also a great opportunity to refer back to your base project and compare. feedback Put it Back Together Once you’ve completed your mini-projects, put everything back together again and do the entire project from scratch. You’ll probably be amazed by how smooth the process is by this point. You’ll also have a much clearer understanding of how the individual parts (the mini-projects) fit together as a whole. 2. Extend It Another suggestion for creating projects is to extend an existing one. Whereas the “whole-part” method breaks down and drills concepts that you’re already familiar with, this one has you add something . The idea is to extend your knowledge. new Say you just completed a backend project. Now extend it by adding middleware, if you've not used middleware before. Then extend it once more by creating a new React app and connecting it to the server. You can extend a project in any number of ways, these are just some examples. Each time, you’re going beyond the original project requirements by adding something new to it. In doing so, you’re pushing yourself outside of your comfort zone just a bit, which is the intent. That’s how you’ll get better. Once you start using these approaches, you’ll find yourself creating your own projects from scratch. You’ll take “x” from this mini-project and “y” from another and create a new project. The process will also get you thinking creatively of entirely new projects you can put together, and it’s a lot of fun! When you find a good project, get the most from it. Hopefully, this article gave you some ideas of how to do that. Reference: Nater, Swen. You Haven't Taught Until They Have Learned: John Wooden's Teaching Principles and Practices. Morgantown, Fitness Information Technology, 2006. Programmer and writer: amymhaddad.com | programmerspyramid.com | I tweet about programming, learning, and productivity @amymhaddad Originally published on amymhaddad.com .

Applying John Wooden's "Whole-Part Method" Approach To Programming

Imagine an industry that has seen little innovation since the Civil War. It may seem unbelievable, but the United States railroad infrastructure is still the same as it was back in the original era. Taking on the industry to improve and plan for the future is Evertrak -- ready for a railroad infrastructure revolution. The company's mission is bold: replace traditional wooden railroad ties with cutting-edge, sustainable alternatives made from recycled plastics. Evertrak is addressing environmental concerns and industry challenges head-on. \Tim Noonan, CEO of Evertrak, doesn’t mince words: “It’s 2024!” he emphasizes. “We can do better. We must do better.” His vision of “the railroads of our Nation rolling on ties made from recycled plastics” isn’t just about sustainability—it’s about transforming an industry stuck in the past and solving real problems. Evertrak's journey began with a revolutionary idea: What if the plastic waste choking our planet could become the foundation of our transportation infrastructure? The team realized the industry needed more than just a new product—it needed a paradigm shift. “Our insight was that the industry needed a tie that was as strong as wood, not as cheap as wood,” Noonan explains. “We’re not offering a band-aid. We’re providing a solution for a real problem in the industry.” Addressing Industry Challenges Safety and efficiency are paramount in the railroad industry. With the adoption of Precision Schedule Railroading (PSR), which focuses on moving individual cars rather than waiting for long trains to be built, uptime has become more critical than ever. Evertrak’s ties directly address this need and offer a redefinition of what’s possible in railroad operations. “Using our tie, railroads won't have to replace ties in high decay zones every 8-12 years,” says Noonan. “This means better uptime for PSR and lower total cost of ownership.” Sustainability Impact Experts suggest only 9% of plastic waste is effectively recycled globally. And industry estimates for North America alone suggest 20 million wooden railroad ties are replaced annually—equivalent to approximately 6 million trees. As Noonan mentioned, with an average lifespan of around a decade (reduced to just 5 years in high-rot regions), the environmental impact of current railroad practices is significant and the environmental benefits of Evertrak's ties are clear. By using recycled plastics, Evertrak not only diverts hundreds of thousands of pounds of waste from landfills but also prevents deforestation and avoids the CO2 produced when traditional ties are incinerated or landfilled at the end of their life. “This is just the beginning,” Noonan promises, explaining that the industry replaces more millions of wood ties a year. For Noonan and Evertrak, it’s about securing a livable planet for future generations. “Imagine that many trees saved. Imagine millions of pounds of plastic put to better use and removed from our landfills and oceans. Every tie we produce is a blow against deforestation and plastic pollution.” Union Pacific’s commitment to alternative railroad ties strikes Noonan as a resounding endorsement of this important industry shift, and the CEO sees it as a rallying cry: “UP is the de facto leader of the railroad industry. Their measured implementation of our ties builds trust and confidence for others. I think the message is: It’s time to get on board.” Industry Veterans Signal Transformation One of the most promising signs of Evertrak’s potential to revolutionize the railroad industry is the company’s ability to attract respected executives who have historically been on the wood tie side of the business. “Yes, it really signals disruption in the industry,” Noonan explains. “These executives understand that the Class 1 railroads do not have alternatives when it comes to ties. They understand the business practices and products that keep things going between big creosote and the railroads.” This insider knowledge is proving invaluable as Evertrak seeks to transform the industry. With the never-ending rise of single-use plastic and the detrimental impact of deforestation, Noonan believes there's room for improvement. “We can do better,” he asserts. “I've been speaking with a number of industry veterans who believe that the time is right for transformation.” The involvement of these industry veterans not only lends credibility to Evertrak’s mission but also suggests that the railroad industry as a whole might be gearing up for a significant shift towards more sustainable practices. Future of Sustainable Infrastructure in Railroads Evertrak’s highest priority is continuing to build trust with Class 1 customers, bringing awareness to one of America’s oldest industries—some of the first tracks were laid nearly 200 years ago—so they have the resources and support to join the modern era. The company’s goals are ambitious: save millions of trees, drastically reduce plastic buildup, and do it all now. Evertrak’s approach is collaborative, empowering the customer to dictate success through communication and keeping commitments. “We have 7 years of proven results,” Noonan says. “And that includes never missing a delivery timeline. To date, we’ve installed over 100,000 composite ties with zero claims.” For other companies looking to innovate with sustainable materials, Noonan offers this advice: “Identify a pain point and make that pain go away for the customer. Your customer doesn’t care that the product is made from sustainable material; they want their problem to go away. Deliver the business case, and the case for sustainability comes along to close the deal.” As the railroad industry grapples with the need for more sustainable practices, companies like Evertrak are proving that environmental responsibility and business efficiency can go hand in hand. By addressing real industry pain points with innovative, sustainable solutions, Evertrak is helping to lay the tracks for a sustainable future in transportation infrastructure. The implications of this shift are far-reaching, potentially influencing other sectors of the transportation industry. As Noonan puts it, "If a pain point in the business can be solved through a sustainable solution, I believe the industry can have the business case and the benefit of sustainability without a tradeoff." With industry leaders like Union Pacific embracing these innovations, it's clear that the future of railroad infrastructure isn’t about maintaining the status quo—it’s about revolutionizing it for a more sustainable, efficient tomorrow. In Noonan’s words, “It’s time for the railroad industry to declare its independence from wood ties and creosote. Join us in this revolution! We’re not just making ties—we’re making history.” Imagine an industry that has seen little innovation since the Civil War. It may seem unbelievable, but the United States railroad infrastructure is still the same as it was back in the original era. Taking on the industry to improve and plan for the future is Evertrak -- ready for a railroad infrastructure revolution. The company's mission is bold: replace traditional wooden railroad ties with cutting-edge, sustainable alternatives made from recycled plastics. Evertrak is addressing environmental concerns and industry challenges head-on. Evertrak Evertrak \Tim Noonan, CEO of Evertrak, doesn’t mince words: “It’s 2024!” he emphasizes. “We can do better. We must do better.” His vision of “the railroads of our Nation rolling on ties made from recycled plastics” isn’t just about sustainability—it’s about transforming an industry stuck in the past and solving real problems. “It’s 2024!” he emphasizes. “We can do better. We must do better.” His vision of “the railroads of our Nation rolling on ties made from recycled plastics” isn’t just about sustainability—it’s about transforming an industry stuck in the past and solving real problems. must sustainability—it’s about transforming an industry stuck in the past and solving real problems. Evertrak's journey began with a revolutionary idea: What if the plastic waste choking our planet could become the foundation of our transportation infrastructure? The team realized the industry needed more than just a new product—it needed a paradigm shift. “Our insight was that the industry needed a tie that was as strong as wood, not as cheap as wood,” Noonan explains. “We’re not offering a band-aid. We’re providing a solution for a real problem in the industry.” Addressing Industry Challenges Addressing Industry Challenges Safety and efficiency are paramount in the railroad industry. With the adoption of Precision Schedule Railroading (PSR), which focuses on moving individual cars rather than waiting for long trains to be built, uptime has become more critical than ever. Evertrak’s ties directly address this need and offer a redefinition of what’s possible in railroad operations. “Using our tie, railroads won't have to replace ties in high decay zones every 8-12 years,” says Noonan. “This means better uptime for PSR and lower total cost of ownership.” Sustainability Impact Sustainability Impact Experts suggest only 9% of plastic waste is effectively recycled globally. And industry estimates for North America alone suggest 20 million wooden railroad ties are replaced annually—equivalent to approximately 6 million trees. As Noonan mentioned, with an average lifespan of around a decade (reduced to just 5 years in high-rot regions), the environmental impact of current railroad practices is significant and the environmental benefits of Evertrak's ties are clear. only 9% only 9% By using recycled plastics, Evertrak not only diverts hundreds of thousands of pounds of waste from landfills but also prevents deforestation and avoids the CO2 produced when traditional ties are incinerated or landfilled at the end of their life. “This is just the beginning,” Noonan promises, explaining that the industry replaces more millions of wood ties a year. For Noonan and Evertrak, it’s about securing a livable planet for future generations. “Imagine that many trees saved. Imagine millions of pounds of plastic put to better use and removed from our landfills and oceans. Every tie we produce is a blow against deforestation and plastic pollution.” that millions Union Pacific’s commitment to alternative railroad ties strikes Noonan as a resounding endorsement of this important industry shift, and the CEO sees it as a rallying cry: “UP is the de facto leader of the railroad industry. Their measured implementation of our ties builds trust and confidence for others. I think the message is: It’s time to get on board.” Industry Veterans Signal Transformation One of the most promising signs of Evertrak’s potential to revolutionize the railroad industry is the company’s ability to attract respected executives who have historically been on the wood tie side of the business. “Yes, it really signals disruption in the industry,” Noonan explains. “These executives understand that the Class 1 railroads do not have alternatives when it comes to ties. They understand the business practices and products that keep things going between big creosote and the railroads.” “Yes, it really signals disruption in the industry,” Noonan explains. “These executives understand that the Class 1 railroads do not have alternatives when it comes to ties. They understand the business practices and products that keep things going between big creosote and the railroads.” This insider knowledge is proving invaluable as Evertrak seeks to transform the industry. With the never-ending rise of single-use plastic and the detrimental impact of deforestation, Noonan believes there's room for improvement. “We can do better,” he asserts. “I've been speaking with a number of industry veterans who believe that the time is right for transformation.” The involvement of these industry veterans not only lends credibility to Evertrak’s mission but also suggests that the railroad industry as a whole might be gearing up for a significant shift towards more sustainable practices. Future of Sustainable Infrastructure in Railroads Evertrak’s highest priority is continuing to build trust with Class 1 customers, bringing awareness to one of America’s oldest industries—some of the first tracks were laid nearly 200 years ago—so they have the resources and support to join the modern era. The company’s goals are ambitious: save millions of trees, drastically reduce plastic buildup, and do it all now. Evertrak’s approach is collaborative, empowering the customer to dictate success through communication and keeping commitments. “We have 7 years of proven results,” Noonan says. “And that includes never missing a delivery timeline. To date, we’ve installed over 100,000 composite ties with zero claims.” zero For other companies looking to innovate with sustainable materials, Noonan offers this advice: “Identify a pain point and make that pain go away for the customer. Your customer doesn’t care that the product is made from sustainable material; they want their problem to go away. Deliver the business case, and the case for sustainability comes along to close the deal.” As the railroad industry grapples with the need for more sustainable practices, companies like Evertrak are proving that environmental responsibility and business efficiency can go hand in hand. By addressing real industry pain points with innovative, sustainable solutions, Evertrak is helping to lay the tracks for a sustainable future in transportation infrastructure. The implications of this shift are far-reaching, potentially influencing other sectors of the transportation industry. As Noonan puts it, "If a pain point in the business can be solved through a sustainable solution, I believe the industry can have the business case and the benefit of sustainability without a tradeoff." With industry leaders like Union Pacific embracing these innovations, it's clear that the future of railroad infrastructure isn’t about maintaining the status quo—it’s about revolutionizing it for a more sustainable, efficient tomorrow. In Noonan’s words, “It’s time for the railroad industry to declare its independence from wood ties and creosote. Join us in this revolution! We’re not just making ties—we’re making history.”

Railroad Infrastructure: How Evertrak is Leading the Way with Sustainable Materials

When I first transitioned to remote work, I underestimated the importance of having a dedicated workspace, relying solely on my laptop and couch. However, this approach soon led to back and neck pain and a disrupted work-life balance, making me realize the need to create a tailored working environment in my apartment. Here’s what I’ve learned from organizing my home office, where I’ve been working for almost four years now. A home office should be both functional and visually pleasing. Otherwise, you risk experiencing a decline in productivity and your overall sense of well-being. Arrange your work from the home area in a minimalist style, including only the furniture and decor items necessary for physical and mental comfort and functional tasks. Excessive and unfitting stuff can mess up your workspace and make you feel overwhelmed and tense. Another big deal is keeping your home office private. Choose a location that minimizes distractions. A separate room with a door is ideal, but if your apartment planning doesn’t allow this, find a quiet corner or use room dividers. Below, I’ll share five practical tips for organizing an effective workspace and setting an advantageous atmosphere for your remote job. Prioritize Furnishing Essentials Your top priority is maintaining your body’s comfort during long workdays. Invest in an ergonomic chair with good lumbar support, adjustable height, and armrests, and a writing desk spacious enough to accommodate your computer and other work essentials you may need. I prefer shopping for furniture offline to ensure the items meet my needs. For example, I test the chair for proper spine support and confirm that the desk is tall enough to avoid knocking my knees on its edge. Placing a footrest under your desk will help promote a proper posture while sitting and help you keep a neutral position throughout the workday. For added comfort, you can put a cushion on your chair. If made from soft fabrics, it will promote relaxation and help you return to your body sensations. Create a Favorable Background Place a freestanding or wall-mounted shelving unit behind your work desk. Ideally, choose one that is vertical and matches the width of your desk to avoid taking up too much floor space. Fill the shelves with books and decorative items to create a stylish, professional background infused with your personal touch. This setup creates a visually appealing backdrop for virtual meetings and ensures that you’ll always be ready for calls, including unplanned ones, in your neat, small office, even when the rest of your apartment is messy during a busy week. Add a Nature Touch As with your living space, your workspace shouldn't be packed with artificial-only items, making it a nature-depravated zone and negatively affecting your mental state. It will also drastically drop the air quality in your room, causing stuffiness and decreasing concentration. The lifeful decor elements such as plants, wooden furniture, aquariums, imitation waterfalls, and more can fulfill your need for affinity to the natural world in your working environment, boosting your enthusiasm and productivity. Be Mindful of Colors Choosing a color scheme that resonates with your personal preferences is no less important than getting comfortable furniture. Gloomy color solutions or flashy-bright design patterns act as ultimate stressors for your mind, overloading your eyes and brain and draining your productivity without you even realizing it. Overall, a natural base of whites, grays, or beiges will bring a tranquil sounding to your work surroundings. Pastel colors promote serenity and comfort, and warm neutrals are uplifting. You can also incorporate accent colors, like blues and greens for focus and concentration, and yellows or oranges for energy and creativity. Ensure Adequate Lighting Position your desk near a window to take advantage of natural light, which is essential for maintaining the body’s healthy internal clock. Exposure to sunlight helps regulate sleep patterns and prevent chronic fatigue. In contrast, working in poorly lit environments can increase stress and anxiety, cause energy dips, and reduce productivity and motivation. If you work late in the evenings, invest in a desk lamp with adjustable brightness and ensure the room is well-lit with overhead lighting to avoid eye strain. Last but not least, you should regularly clean and declutter your home office compartment and keep your work essentials in order. A tidy environment enhances focus and productivity, while a not-properly-organized space can be disorienting and provoke procrastination. When I first transitioned to remote work, I underestimated the importance of having a dedicated workspace, relying solely on my laptop and couch. However, this approach soon led to back and neck pain and a disrupted work-life balance, making me realize the need to create a tailored working environment in my apartment. Here’s what I’ve learned from organizing my home office, where I’ve been working for almost four years now. When I first transitioned to remote work, I underestimated the importance of having a dedicated workspace, relying solely on my laptop and couch. However, this approach soon led to back and neck pain and a disrupted work-life balance, making me realize the need to create a tailored working environment in my apartment. Here’s what I’ve learned from organizing my home office, where I’ve been working for almost four years now. A home office should be both functional and visually pleasing. Otherwise, you risk experiencing a decline in productivity and your overall sense of well-being. Arrange your work from the home area in a minimalist style, including only the furniture and decor items necessary for physical and mental comfort and functional tasks. Excessive and unfitting stuff can mess up your workspace and make you feel overwhelmed and tense. Another big deal is keeping your home office private. Choose a location that minimizes distractions. A separate room with a door is ideal, but if your apartment planning doesn’t allow this, find a quiet corner or use room dividers. Below, I’ll share five practical tips for organizing an effective workspace and setting an advantageous atmosphere for your remote job. Prioritize Furnishing Essentials Your top priority is maintaining your body’s comfort during long workdays. Invest in an ergonomic chair with good lumbar support, adjustable height, and armrests, and a writing desk spacious enough to accommodate your computer and other work essentials you may need. I prefer shopping for furniture offline to ensure the items meet my needs. For example, I test the chair for proper spine support and confirm that the desk is tall enough to avoid knocking my knees on its edge. Placing a footrest under your desk will help promote a proper posture while sitting and help you keep a neutral position throughout the workday. For added comfort, you can put a cushion on your chair. If made from soft fabrics, it will promote relaxation and help you return to your body sensations. Create a Favorable Background Place a freestanding or wall-mounted shelving unit behind your work desk. Ideally, choose one that is vertical and matches the width of your desk to avoid taking up too much floor space. Fill the shelves with books and decorative items to create a stylish, professional background infused with your personal touch. This setup creates a visually appealing backdrop for virtual meetings and ensures that you’ll always be ready for calls, including unplanned ones, in your neat, small office, even when the rest of your apartment is messy during a busy week. Add a Nature Touch As with your living space, your workspace shouldn't be packed with artificial-only items, making it a nature-depravated zone and negatively affecting your mental state. It will also drastically drop the air quality in your room, causing stuffiness and decreasing concentration. The lifeful decor elements such as plants, wooden furniture, aquariums, imitation waterfalls, and more can fulfill your need for affinity to the natural world in your working environment, boosting your enthusiasm and productivity. Be Mindful of Colors Choosing a color scheme that resonates with your personal preferences is no less important than getting comfortable furniture. Gloomy color solutions or flashy-bright design patterns act as ultimate stressors for your mind, overloading your eyes and brain and draining your productivity without you even realizing it. Overall, a natural base of whites, grays, or beiges will bring a tranquil sounding to your work surroundings. Pastel colors promote serenity and comfort, and warm neutrals are uplifting. You can also incorporate accent colors, like blues and greens for focus and concentration, and yellows or oranges for energy and creativity. Ensure Adequate Lighting Position your desk near a window to take advantage of natural light, which is essential for maintaining the body’s healthy internal clock. Exposure to sunlight helps regulate sleep patterns and prevent chronic fatigue. In contrast, working in poorly lit environments can increase stress and anxiety, cause energy dips, and reduce productivity and motivation. If you work late in the evenings, invest in a desk lamp with adjustable brightness and ensure the room is well-lit with overhead lighting to avoid eye strain. Last but not least, you should regularly clean and declutter your home office compartment and keep your work essentials in order. A tidy environment enhances focus and productivity, while a not-properly-organized space can be disorienting and provoke procrastination.

5 Home Office Setup Ideas for Productivity and Work-Life Balance

Everything You Need to Know About Trump v. United States

Many organizations that we’ve spoken to are in the exploration phase of using for AI-powered personalization, recommendations, semantic search and anomaly detection. The recent and astronomical improvements in accuracy and accessibility of large language models (LLMs) including BERT and OpenAI have made companies rethink how to build relevant search and analytics experiences. vector search In this blog, we capture engineering stories from 5 early adopters of vector search- Pinterest, Spotify, eBay, Airbnb and Doordash- who have integrated AI into their applications. We hope these stories will be helpful to engineering teams who are thinking through the full lifecycle of vector search all the way from generating embeddings to production deployments. What is vector search? Vector search is a method for efficiently finding and retrieving similar items from a large dataset based on representations of the data in a high-dimensional space. In this context, items can be anything, such as documents, images, or sounds, and are represented as vector embeddings. The similarity between items is computed using distance metrics, such as or , which quantify the closeness of two vector embeddings. cosine similarity Euclidean distance The vector search process usually involves: : Where relevant features are extracted from the raw data to create vector representations using models such as , or Generating embeddings word2vec BERT Universal Sentence Encoder : The vector embeddings are organized into a data structure that enables efficient search using algorithms such as or Indexing FAISS HNSW : Where the most similar items to a given query vector are retrieved based on a chosen distance metric like cosine similarity or Euclidean distance Vector search To better visualize vector search, we can imagine a 3D space where each axis corresponds to a feature. The time and the position of a point in the space is determined by the values of these features. In this space, similar items are located closer together and dissimilar items are farther apart. GITHUB JULIE-MILLS ^
  |          x Item 1
  |         /
  |        /
  |       /x Item 2
  |      /
  |     /
  |    /x Item 3
  |   /
  |  /
  | /
  +-------------------> Given a query, we can then find the most similar items in the dataset. The query is represented as a vector embedding in the same space as the item embeddings, and the distance between the query embedding and each item embedding is computed. The item embeddings with the shortest distance to the query embedding are considered the most similar. Query item:         x
                    |
  ^                 |
  |          x Item 1|
  |         /       |
  |        /        |
  |       /x Item 2 |
  |      /          |
  |     /           |
  |    /x Item 3    |
  |   /             |
  |  /              |
  | /               |
  +-------------------> This is obviously a simplified visualization as vector search operates in high-dimensional spaces. In the next sections, we’ll summarize 5 engineering blogs on vector search and highlight key implementation considerations. The full engineering blogs can be found below: by Jinfeng Zhuang at Pinterest PinText: A Multitask Text Embedding System in Pinterest by Alexandre Tamborrino at Spotify Introducing Natural Language Search for Podcast Episodes by Senthilkumar Gopal, Shubhangi Tandon, Christopher Miller, Deepika Srinivasan, Rui Kong, Selcuk Kopru and Srinivas Bhagavathula at eBay How eBay’s New Search Feature Was Inspired By Window Shopping by Mihajlo Grbovic at AirBnb Listing Embeddings in Search Ranking by Mitchell Koch, Aamir Manasawala, Raghav Ramesh at Doordash Personalized Store Feed with Vector Embeddings Pinterest: Interest search and discovery Pinterest uses for image search and discovery across multiple areas of its platform, including recommended content on the home feed, related pins and search using a multitask learning model. vector search A multi-task model is trained to perform multiple tasks simultaneously, often sharing underlying representations or features, which can improve generalization and efficiency across related tasks. In the case of Pinterest, the team trained and used the same model to drive recommended content on the homefeed, related pins and search. Pinterest trains the model by pairing a users search query (q) with the content they clicked on or pins they saved (p). Here is how Pinterest created the (q,p) pairs for each task: : Word embeddings are derived from the selected subject (q) and the pin clicked on or saved by the user (p). Related Pins : Word embeddings are created from the search query text (q) and the pin clicked on or saved by the user (p). Search : Word embeddings are generated based on the interest of the user (q) and the pin clicked on or saved by the user (p). Homefeed To obtain an overall entity embedding, Pinterest averages the associated word embeddings for related pins, search and the homefeed. Pinterest created and evaluated its own supervised Pintext-MTL (multi-task learning) against unsupervised learning models including GloVe, word2vec as well as a single-task learning model, PinText-SR on precision. PinText-MTL had higher precision than the other embedding models, meaning that it had a higher proportion of true positive predictions among all positive predictions. Pinterest also found that multi-task learning models had a higher recall, or a higher proportion of relevant instances correctly identified by the model, making them a better fit for search and discovery. To put this all together in production, Pinterest has a multitask model trained on streaming data from the homefeed, search and related pins. Once that model is trained, vector embeddings are created in a large batch job using either Kubernetes+Docker or a map-reduce system. The platform builds a search index of vector embeddings and runs a K-nearest neighbors (KNN) search to find the most relevant content for users. Results are cached to meet the performance requirements of the Pinterest platform. Spotify: Podcast search Spotify combines keyword and to retrieve relevant podcast episode results for users. As an example, the team highlighted the limitations of keyword search for the query “electric cars climate impact”, a query which yielded 0 results even though relevant podcast episodes exist in the Spotify library. To improve recall, the Spotify team used Approximate Nearest Neighbor (ANN) for fast, relevant podcast search. semantic search The team generates vector embeddings using the as it is multilingual, supporting a global library of podcasts, and produces high-quality vector embeddings. Other models were also evaluated including , a model trained on a big corpus of text data, but found that BERT was better suited for word embeddings than sentence embeddings and was pre-trained only in English. Universal Sentence Encoder CMLM model BERT Spotify builds the vector embeddings with the query text being the input embedding and a concatenation of textual metadata fields including title and description for the podcast episode embeddings. To determine the similarity, Spotify measured the cosine distance between the query and episode embeddings. To train the base Universal Sentence Encoder CMLM model, Spotify used positive pairs of successful podcast searches and episodes. They incorporated in-batch negatives, a technique highlighted in papers including and , to generate random negative pairings. Testing was also conducted using synthetic queries and manually written queries. Dense Passage Retrieval for Open-Domain Question Answering (DPR) Que2Search: Fast and Accurate Query and Document Understanding for Search at Facebook To incorporate vector search into serving podcast recommendations in production, Spotify used the following steps and technologies: : Spotify indexes the episode vectors offline in batch using , a search engine with native support for ANN. One of the reasons that Vespa was chosen is that it can also incorporate metadata filtering post-search on features like episode popularity. Index episode vectors Vespa : Spotify uses to generate a query vector. Vertex AI was chosen for its support for GPU inference, which is more cost effective when using large transformer models to generate embeddings, and for its query cache. After the query vector embedding is generated, it is used to retrieve the top 30 podcast episodes from Vespa. Online inference Google Cloud Vertex AI Semantic search contributes to the identification of pertinent podcast episodes, yet it is unable to fully supplant keyword search. This is due to the fact that semantic search falls short of exact term matching when users search an exact episode or podcast name. Spotify employs a hybrid search approach, merging semantic search in Vespa with keyword search in , followed by a conclusive re-ranking stage to establish the episodes displayed to users. Elasticsearch eBay: Image search Traditionally, search engines have displayed results by aligning the search query text with textual descriptions of items or documents. This method relies extensively on language to infer preferences and is not as effective in capturing elements of style or aesthetics. eBay introduces to help users find relevant, similar items that meet the style they’re looking for. image search eBay uses a multi-modal model which is designed to process and integrate data from multiple modalities or input types, such as text, images, audio, or video, to make predictions or perform tasks. eBay incorporates both text and images into its model, producing image embeddings utilizing a Convolutional Neural Network (CNN) model, specifically , and title embeddings using a text-based model such as . Every listing is represented by a vector embedding that combines both the image and title embeddings. Resnet-50 BERT Once the multi-modal model is trained using a large dataset of image-title listing pairs and recently sold listings, it is time to put it into production in the site search experience. Due to the large number of listings at eBay, the data is loaded in batches to HDFS, eBay’s data warehouse. eBay uses Apache Spark to retrieve and store the image and relevant fields required for further processing of listings, including generating listing embeddings. The listing embeddings are published to a columnar store such as HBase which is good at aggregating large-scale data. From HBase, the listing embedding is indexed and served in Cassini, a search engine created at eBay. The pipeline is managed using Apache Airflow, which is capable of scaling even when there is a high quantity and complexity of tasks. It also provides support for Spark, Hadoop, and Python, making it convenient for the machine learning team to adopt and utilize. Visual search allows users to find similar styles and preferences in the categories of furniture and home decor, where style and aesthetics are key to purchase decisions. In the future, eBay plans to expand visual search across all categories and also help users discover related items so they can establish the same look and feel across their home. AirBnb: Real-time personalized listings Search and similar listings features drive 99% of bookings on the AirBnb site. AirBnb built a to improve similar listing recommendations and provide real-time personalization in search rankings. listing embedding technique AirBnb realized early on that they could expand the application of embeddings beyond just word representations, encompassing user behaviors including clicks and bookings as well. To train the embedding models, AirBnb incorporated over 4.5M active listings and 800 million search sessions to determine the similarity based on what listings a user clicks and skips in a session. Listings that were clicked by the same user in a session are pushed closer together; listings that were skipped by the user are pushed further away. The team settled on the dimensionality of a listing embedding of d=32 given the tradeoff between offline performance and memory needed for online serving. https://youtu.be/aWjsUEX7B1I?si=GREROVrWxWqtqL-m&embedable=true AirBnb found that certain listings characteristics do not require learning, as they can be directly obtained from metadata, such as price. However, attributes like architecture, style, and ambiance are considerably more challenging to derive from metadata. Before moving to production, AirBnb validated their model by testing how well the model recommended listings that a user actually booked. The team also ran an A/B test comparing the existing listings algorithm against the vector embedding-based algorithm. They found that the algorithm with vector embeddings resulted in a 21% uptick in CTR and 4.9% increase in users discovering a listing that they booked. The team also realized that vector embeddings could be used as part of the model for real-time personalization in search. For each user, they collected and maintained in real time, using Kafka, a short-term history of user clicks and skips in the last two weeks. For every search conducted by the user, they ran two similarity searches: based on the geographic markets that were recently searched and then the similarity between the candidate listings and the ones the user has clicked/skipped Embeddings were evaluated in offline and online experiments and became part of the real-time personalization features. Doordash: Personalized store feeds Doordash has a wide variety of stores that users can choose to order from and being able to surface the most relevant stores using personalized preferences improves search and discovery. Doordash wanted to apply to its store feed algorithms using vector embeddings. This would enable Doordash to uncover similarities between stores that were not well-documented including if a store has sweet items, is considered trendy or features vegetarian options. latent information Doordash used a derivative of word2vec, an embedding model used in natural language processing, called store2vec that it adapted based on existing data. The team treated each store as a word and formed sentences using the list of stores viewed during a single user session, with a maximum limit of 5 stores per sentence. To create user vector embeddings, Doordash summed the vectors of the stores from which users placed orders in the past 6 months or up to 100 orders. As an example, Doordash used vector search to find similar restaurants for a user based on their recent purchases at popular, trendy joints 4505 Burgers and New Nagano Sushi in San Francisco. Doordash generated a list of similar restaurants measuring the cosine distance from the user embedding to store embeddings in the area. You can see that the stores that were closest in cosine distance include Kezar Pub and Wooden Charcoal Korean Village BBQ. Doordash incorporated store2vec distance feature as one of the features in its larger recommendation and personalization model. With vector search, Doordash was able to see a 5% increase in click-through-rate. The team is also experimenting with new models like , model optimizations and incorporating real-time onsite activity data from users. seq2seq Key considerations for vector search Pinterest, Spotify, eBay, Airbnb and Doordash create better search and discovery experiences with vector search. Many of these teams started out using text search and found limitations with fuzzy search or searches of specific styles or aesthetics. In these scenarios, adding vector search to the experience made it easier to find relevant, and often personalized, podcasts, pillows, rentals, pins and eateries. There are a few decisions that these companies made that are worth calling out when implementing vector search: : Many started out using an off-the-shelf model and then trained it on their own data. They also recognized that language models like word2vec could be used by swapping words and their descriptions with items and similar items that were recently clicked. Teams like AirBnb found that using derivatives of language models, rather than image models, could still work well for capturing visual similarities and differences. Embedding models : Many of these companies opted to train their models on past purchase and click through data, making use of existing large-scale datasets. Training : While many companies adopted ANN search, we saw that Pinterest was able to combine metadata filtering with KNN search for efficiency at scale. Indexing : Vector search rarely replaces text search. Many times, like in Spotify’s example, a final ranking algorithm is used to determine whether vector search or text search generated the most relevant result. Hybrid search : We’re seeing many teams use batch-based systems to create the vector embeddings, given that these embeddings are rarely updated. They employ a different system, frequently Elasticsearch, to compute the query vector embedding live and incorporate real-time metadata in their search. Productionizing Rockset, a real-time search and analytics database, recently added support for . Give vector search on Rockset a try for real-time personalization, recommendations, anomaly detection and more by starting a with $300 in credits today. vector search free trial

WOEDEN

woeden.com

-- employees

Since n.d.

COMPANY RANKING

WOEDEN

EVERGREEN INDEX #11721

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

Woeden WIKI

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

WOEDEN

woeden.com

-- employees

Since n.d.

COMPANY RANKING

WOEDEN

EVERGREEN INDEX #11721

RELATED COMPANIESRANK

STORIES ABOUT WOEDEN

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

Woeden WIKI

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