Lynx Roundup, February 19th 2020 Game theory and reinforcement learning! Digital Socialism! Feature engineer optimization in HyperparameterHunter! Matthew Alhonte 19 Feb 2020 This Obscure Area of Game Theory can Help to Scale Reinforcement Learning to Infinite AgentsReinforcement learning is one of the most popular areas of research in deep learning nowadays. Part of the popularity of reinforcement learning is due to the fact that is one of the learning methods…Towards Data ScienceJesus RodriguezTry out your climate policy ideasThe Energy Policy Simulator is Canada’s first free, open-source tool giving you a chance to see how energy and climate policies influence emissions across the country.Pembina InstitutePembina InstituteMathematical framework turns any sheet of material into any shape using kirigami cutsResearchers from the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) have developed a mathematical framework that can turn any sheet of material into any prescribed shape, inspired by the paper craft termed kirigami (from the Japanese, kiri, meaning to cut and kami, meaning…Phys.orgScience X staffEvgeny Morozov, Digital Socialism?, NLR 116/117, March–June 2019What opportunities does the new feedback infrastructure of data capitalism offer for non-market forms of social coordination? Hayek and the Calculation Debate revisited in the age of big data.New Left Review‘Socialize the Data Centres!’ In today’s post, I am looking at Wittgenstein’s ladder at the gemba. Ludwig Wittgenstein is one of the most profound philosophers of the 20th century. His first book was Tractatus Logico-Philosophicus, in which he came up with the picture theory of language. He defined how language and reality relate to each other, and how limits of language corresponded to limits of knowledge to some extent. Loosely put, the Tractatus explained how language can be used to directly depict reality. Language should mirror exactly the arrangement of objects, and their relationships to each other in the real world. Wittgenstein proposed that what can be said about the world makes sense only if there is a correspondence to the real world out there. Everything else is nonsense. This idea puts limits to how we use language. The real use of language is to describe reality. Anthony Quinton, the late British philosopher, explained the main concepts of Tractatus as: Tractatus is a theory of declarative sentences, a theory of what can be put in a proposition and what cannot. Anything that can be said can be said clearly or not at all. The world is all that is the case. The state of affairs around us, the simple facts, are the world for us. Wittgenstein is talking about what we can and cannot sensibly talk about. The world consists of facts. Facts are arrangement of objects. Objects must be simple. These ideas appear as dogmatic assertions. Language has to have a definite sense and it can have a definite sense only if it is of a certain structure. And therefore the world must be of that certain structure in order to be capable of being represented in the language. One of the metaphors, Wittgenstein used in the Tractatus is the idea of a ladder. This has come to be known as “Wittgenstein’s Ladder.” Wittgenstein said: My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them—as steps—to climb beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.) He must transcend these propositions, and then he will see the world aright. This is a fascinating idea because Wittgenstein is cautioning against doctrines as the eternal rules to abide by. If the concepts that Wittgenstein explained in the Tractatus are true, then the assertion of his ideas being true would contradict the ideas themselves. Wittgenstein uses the metaphor of a ladder to have the reader climb to a higher level of understanding and then asks the reader to kick the ladder away. Let’s see how Wittgenstein’s ladder relates to Lean/Toyota Production System. Taiichi Ohno developed TPS as a production system through decades of trial and error methods. The solutions Ohno came up with were specific to the problems Toyota had at that time. We should learn about these different tools and understand the problems they are trying to solve. We should not exactly copy the tools that Toyota uses just because Toyota is using them. Even within Toyota, each plant is unique and doesn’t use a specific set of tools. As one Toyota veteran put it, Toyota Production System and Toyota’s Production System are different. What each plant does is unique and based on the complexity of problems it has. There are several doctrines that are set forth by the experts. Let’s look at two examples – zero inventories and one-piece flow. Taiichi Ohno himself tried to correct these two misrepresentations/misunderstandings. Ohno called the Zero Inventory idea nonsense: To be sure, if we completely eliminate inventories, we will have shortages of goods and other problems. In fact, reducing inventories to zero is nonsense. The goal of Toyota Production System is to level the flows of production and goods… In every plant and retail outlet, we strive to have the needed goods arrive in the needed quantities in the needed time. In no way is the Toyota Production System a zero-inventory system. Similarly, Ohno also cautioned about implementing one-piece flow without thinking and looking at your production system. The essence of Toyota Production System is found in the saying, “Can we realistically reduce one more?” and then after that “one more?” The removal of parts or operators is about identifying waste and ways to improve human capital through problem solving. The idea is to develop people and not think only about developing parts. Kaizen is a philosophy of personal improvement (improving oneself) through process improvements. Kaizen begets more kaizen. Final Words: The problem with doctrines is that we build a religion out of them. Ask yourself – What is the problem that I am trying to solve? Toyota’s solutions work for Toyota’s problems. We should climb the TPS/Lean ladder (understand the ideas) and then throw away the ladder of doctrines. We should solve our problems using solutions that match our problems. Always keep on learning… In case you missed it, my last post was Drawing at the Gemba: Learning in Graphs with Python (Part 3)Graphs are becoming central to machine learning these days, whether you’d like to understand the structure of a social network by predicting potential connections, detecting fraud, understand…Towards Data ScienceMaël FabienFeature Engineer Optimization in HyperparameterHunter 3.0Automatically save and optimize your feature engineering steps alongside hyperparameters with HyperparameterHunter — making optimization smarter and ensuring no Experiment is wasted The long wait is…Towards Data ScienceHunter McGushion Roundup Matthew Alhonte @MattAlhonte Supervillain in somebody's action hero movie. Experienced a radioactive freak accident at a young age which rendered him part-snake and strangely adept at Python. Hackers and Slackers Newsletter Join the newsletter to receive the latest updates in your inbox. Your email address Subscribe Please check your inbox and click the link to confirm your subscription. Please enter a valid email address! An error occurred, please try again later.