ConfPad

"Methods that scale with computation are the future of AI", Richard Sutton, father of reinforcement learning. Large labelled training datasets were only one of the key pillars of the deep learning revolution, the widespread availability of GPU compute was the other. The next phase of deep learning is the widespread availability of distributed GPU compute. As…

Eliot explains how to use deep learning models in near real-time with a large amount of data, taking as an example Linkfluence who is processing more than 100 million images per month. He details the general architecture, pitfalls to avoid as well as some tips and tricks.

Machine Learning on Source Code (#MLonCode) is a new cool domain of Machine Learning which takes source code as an input data. Vadim tells the story about capturing the code naturalness, one of the core concepts in MLonCode, through identifier (class, function, variable name) embeddings. Embedding millions of identifiers is challenging... watch how this prob…

In logistics services, the only way to grow is to reduce costs, as this is a cost centre for many managers. On average, in France and in all business sectors, logistics costs represent between 8 and 10% of sales. Data science and associated tools can help reduce these costs while improving customer satisfaction. This means faster delivery, lower costs and fe…

A machine learning instance is composed of: an execution machine, a mathematical model, its knowledge (the weights of the matrices). As a conclusion, he's introducing his Babel Fish: a tool whose goal is to translate mathematical equations (described in Unicode) into computation graph at runtime. For more technical information: go to the Gorgonia repository.

During a fictional match between a human and an AI algorithm, Virginie reviews 3 traps to avoid when working with Artificial Intelligence and how to tackle them: the accuracy paradox (when you have imbalanced classes), the choice of the evaluation function in behavioral tasks, and spotting overfitting in behavioral tasks More about those traps: the accuracy …

Like many developers and engineers in the field, Isabel started her journey in computer vision in academia. For some years she was working on image processing for biomedical images in the pre-Deep-Learning-era using handcrafted features. About three years ago she was offered a position at a start-up to apply deep learning on real-life projects. Isabel though…

Aurelien explains how you can combine Knowledge Graphs and Deep Learning to dramatically improve Search & Discovery systems. By using a combination of signals (audiovisual content, title & description and context), it is possible to find the main topics of a video. These topics can then be used to improve recommendations, search, structured browsing, ads, an…

Much of Artificial Intelligence Research is done by a handful of elite researchers. At the same time, hundreds of thousands of really talented developers and engineers stand behind a mental block that they need a lot of advanced skills to be able to contribute to AI Research, or even to use AI in their own work. Sharada attempts to help clarify the myth, and…

AI seems impossible to dissociate from Big Data, usually intended to mean hundreds of thousands of training samples if not more. But what if what's large about your data is the number of features? This setup poses different statistical and computational challenges, and traditional feature selection methods fall short. The field of structured sparsity offers …

Deep Learning has been delivering state of the art results across a growing number of problems and domains. Correspondingly, Deep Learning models are being deployed across a growing number of applications and use cases. Hagay shows us what deploying deep neural networks to production mean, design considerations and challenges for model serving, and how the o…

As AI becomes more and more prevalent, the decisions it makes for us are becoming more and more impactful on our lives and those of others. How can we help people to have trust in the models we're building? The field of Explainable AI focuses on making any machine learning model interpretable by non experts.