Artificial intelligence (AI) is expected to have a monumental impact on society. As such, DeepMind, an AI research company now housed under Google parent company Alphabet, has established a new unit dedicated to answering questions about the effect the technology might have on the way we live. DeepMind Ethics and Society will bring together employees from the company and outsiders who are uniquely equipped to offer useful perspectives. Economist and former UN advisor Jeffrey Sachs, University of Oxford AI professor Nick Bostrom, and climate change campaigner Christiana Figueres are among the advisers selected for the group. At present, the unit comprises around eight DeepMind employees and six unpaid fellows from outside the company.

Follow leaders in ML on twitter to see what research papers/blog posts/etc. This is a very effective but highly under-rated way to get good at ML. Having seen a lot of new Stanford PhD students grow to become great researchers, I can say confidently that replicating others' results (not just reading the papers) is one of the most effective ways to see and make sure you understand the details of the latest algorithms. Many people jump too quickly into trying to invent something new, which is also worth doing, but is actually a slower way to learn and build up your foundation of knowledge. When you do build something new, publish it in a paper or blog post and consider open-sourcing your code, and share it back out with the community! Hopefully this will help you get more feedback from the community, and further accelerate your learning.

We propose a probabilistic model for interpreting gene expression levels that are observed through single-cell RNA sequencing. In the model, each cell has a low-dimensional latent representation. Additional latent variables account for technical effects that may erroneously set some observations of gene expression levels to zero. Conditional distributions are specified by neural networks, giving the proposed model enough flexibility to fit the data well. We use variational inference and stochastic optimization to approximate the posterior distribution. The inference procedure scales to over one million cells, whereas competing algorithms do not. Even for smaller datasets, for several tasks, the proposed procedure outperforms state-of-the-art methods like ZIFA and ZINB-WaVE. We also extend our framework to take into account batch effects and other confounding factors and propose a natural Bayesian hypothesis framework for differential expression that outperforms tradition DESeq2.

The area of Handwritten Signature Verification has been broadly researched in the last decades, but remains an open research problem. The objective of signature verification systems is to discriminate if a given signature is genuine (produced by the claimed individual), or a forgery (produced by an impostor). This has demonstrated to be a challenging task, in particular in the offline (static) scenario, that uses images of scanned signatures, where the dynamic information about the signing process is not available. Many advancements have been proposed in the literature in the last 5-10 years, most notably the application of Deep Learning methods to learn feature representations from signature images. In this paper, we present how the problem has been handled in the past few decades, analyze the recent advancements in the field, and the potential directions for future research.

Convolutional Neural Networks (CNNs) are a popular deep learning architecture widely applied in different domains, in particular in classifying over images, for which the concept of convolution with a filter comes naturally. Unfortunately, the requirement of a distance (or, at least, of a neighbourhood function) in the input feature space has so far prevented its direct use on data types such as omics data. However, a number of omics data are metrizable, i.e., they can be endowed with a metric structure, enabling to adopt a convolutional based deep learning framework, e.g., for prediction. We propose a generalized solution for CNNs on omics data, implemented through a dedicated Keras layer. In particular, for metagenomics data, a metric can be derived from the patristic distance on the phylogenetic tree. For transcriptomics data, we combine Gene Ontology semantic similarity and gene co-expression to define a distance; the function is defined through a multilayer network where 3 layers are defined by the GO mutual semantic similarity while the fourth one by gene co-expression. As a general tool, feature distance on omics data is enabled by OmicsConv, a novel Keras layer, obtaining OmicsCNN, a dedicated deep learning framework. Here we demonstrate OmicsCNN on gut microbiota sequencing data, for Inflammatory Bowel Disease (IBD) 16S data, first on synthetic data and then a metagenomics collection of gut microbiota of 222 IBD patients.

A successful grasp requires careful balancing of the contact forces. Deducing whether a particular grasp will be successful from indirect measurements, such as vision, is therefore quite challenging, and direct sensing of contacts through touch sensing provides an appealing avenue toward more successful and consistent robotic grasping. However, in order to fully evaluate the value of touch sensing for grasp outcome prediction, we must understand how touch sensing can influence outcome prediction accuracy when combined with other modalities. Doing so using conventional model-based techniques is exceptionally difficult. In this work, we investigate the question of whether touch sensing aids in predicting grasp outcomes within a multimodal sensing framework that combines vision and touch. To that end, we collected more than 9,000 grasping trials using a two-finger gripper equipped with GelSight high-resolution tactile sensors on each finger, and evaluated visuo-tactile deep neural network models to directly predict grasp outcomes from either modality individually, and from both modalities together. Our experimental results indicate that incorporating tactile readings substantially improve grasping performance.

Forecasting financial time series using past observations has been a topic of significant interest for obvious reasons. It is well known that while temporal relationships in the data exist, they are difficult to analyze and predict accurately due to the nonlinear trends and noise present in the series. In developing models for forecasting financial data it is desirable that these will be both able to learn nonlinear dependencies in the data as well as have a high noise resistance. Feedforward neural networks have been a popular way of learning the dependencies in the data, by e.g. using multiple inputs from the past to make a prediction for the future time step, see [3]. One downside of classical feedforward neural networks is that a large sample size of data is required to obtain a stable forecasting result. The main focus of this paper is on multivariate time series forecasting, specifically financial time series. In particular, we forecast time series conditional on other, related series.

This paper proposes a general method for improving the structure and quality of sequences generated by a recurrent neural network (RNN), while maintaining information originally learned from data, as well as sample diversity. An RNN is first pre-trained on data using maximum likelihood estimation (MLE), and the probability distribution over the next token in the sequence learned by this model is treated as a prior policy. Another RNN is then trained using reinforcement learning (RL) to generate higher-quality outputs that account for domain-specific incentives while retaining proximity to the prior policy of the MLE RNN. To formalize this objective, we derive novel off-policy RL methods for RNNs from KL-control. The effectiveness of the approach is demonstrated on two applications; 1) generating novel musical melodies, and 2) computational molecular generation. For both problems, we show that the proposed method improves the desired properties and structure of the generated sequences, while maintaining information learned from data.

I started my deep learning journey a few years back. I have learnt a lot in this period. But, even after all these efforts, every Neural network I train provides me with a new experience. If you have tried to train a neural network, you must know my plight! But, through all this time, I have now made a workflow, which I will share with you today.

Having a hero / heroine helps you navigate through the difficult times. You look up to them and then think that the problems you thought were difficult are actually trivial in nature. If people can solve and deliver at a much larger scale, you can too! If you thought learning data science is difficult or deep neural nets is not your cup of tea – look up to the role models who created them. Following these role models provides you a daily inspiration, a motivation to find bigger purpose in life and to achieve it. Role models set goals for you and try to make you as good as they are. In this article, I'll introduce you to a league of ultimate data scientists in the world.