Ask Slater Machine Learning What is a tensor in physics terminology and what’s the difference from a tensor in machine learning and AI? June 24, 2019 by jeremystinson There’s no difference. A tensor is just a very, very generic term for: basically anything. A tensor is a generic term for a way of arranging numbers that generally has some geometric interpretation. A scalar (math speak for a number) is a 0-tensor. A vector (math speak for a list of numbers) is a 1-tensor […] Read more »
Ask Slater Machine Learning How does the ELMo machine learning model work? June 12, 2019 by jeremystinson At its core, ELMo is an RNN model with an LSTM gating setup that’s trained as a bidirectional language model. But that’s not really what you asked. You asked how it works. It’s often really tough to explain how a machine learning model works. Importantly you have to understand that no part of what I […] Read more »
Ask Slater Machine Learning Should we remove duplicates from a data-set while training a Machine Learning algorithm (shallow and/or deep methods)? February 6, 2019 by jeremystinson It…depends. Mostly it depends on what your goals are and what your dataset looks like. There are two big divides here on both sides. Data Structured Data – here, duplicates very much come with the territory. In this situation you’ve also likely got a lot of implicit ambiguity in your problem. Let’s say that you […] Read more »
Ask Slater Should a machine learning beginner go straight for deep learning? September 26, 2018 by jeremystinson Let me be clear: I love deep learning. It has radically improved the scope of problems that machine learning can be practically applied to. I’ve built a company on the back of deep learning and owe quite a lot to it. DO NOT start with deep learning. I’m not saying that you shouldn’t approach deep […] Read more »
Ask Slater Why do machine learning algorithms require large amounts of computer space for all kinds of datasets? August 29, 2018 by jeremystinson Depends a bit on which model type you’re talking about, but for the purposes of illustration I’m going to assume that we’re talking about something like a Logistic Regression Model on top of a tf-idf vector. In this situation, as well as many others, model size is more or less constant with dataset size. In […] Read more »
Ask Slater What is one thing that human intelligence can do but artificial intelligence can’t? July 25, 2018 / October 9, 2025 by jeremystinson Let me tackle this from a slightly more practical perspective. I think the standard issue with this kind of question is that people are very bad at knowing what human intelligence can do and focus on philosophical questions whose answers are unknown both for humans and for AI. A good example of this is the […] Read more »
Ask Slater What are some of the best artificial intelligence/machine learning journals? July 18, 2018 by jeremystinson Machine learning is a very interesting field. Part of the reason for this is that most of the novel work in the field happens outside of journals. Instead, largely because it advances so quickly, it’s a very conference-driven space. The three biggest conferences are NIPS, ICML, and ICLR. There are definitely a few others, but […] Read more »
Ask Slater Is Euclidean distance meaningful for high dimensional data? July 10, 2018 by jeremystinson The short answer is no. At high dimensions, Euclidean distance loses pretty much all meaning. However, it’s not something that’s the fault of Euclidean distance in particular (though there are distance metrics that work better at high dimensions than Euclidean). The main issue is something commonly referred to as the “Curse of Dimensionality”. It’s very […] Read more »
Ask Slater Is it possible to learn without data (or at least one or two samples)—like humans—in AI? May 26, 2018 by jeremystinson Humans do not learn without data, but as you’ve said, they can often learn with very small amounts of data. This is an extremely active area of research typically referred to as “one-shot learning” (learning from one example) or even “zero-shot learning” (learning from zero examples and just the name of the label you’re looking […] Read more »
Ask Slater Given a scenario, how can I identify which machine learning algorithm is best suited for this scenario? May 26, 2018 by jeremystinson You try them and see how well they work. Anything other than experimentation is guesswork. You can make very broad assertions like, “convolutional neural networks work well for image recognition”, but that doesn’t really help you determine which algorithm you should use, and even in this situation there’s no possible way to know ahead of […] Read more »
