Efficient integration of solar energy into the electricity mix depends on a reliable anticipation of its intermittency. A promising approach to forecast the temporal variability of solar irradiance resulting from the cloud cover dynamics, is based on the analysis of sequences of ground-taken sky images. Despite encouraging results, a recurrent limitation of current Deep Learning approaches lies in the ubiquitous tendency of reacting to past observations rather than actively anticipating future events. This leads to a systematic temporal lag and little ability to predict sudden events. To address this challenge, we introduce ECLIPSE, a spatio-temporal neural network architecture that models cloud motion from sky images to predict both future segmented images and corresponding irradiance levels. We show that ECLIPSE anticipates critical events and considerably reduces temporal delay while generating visually realistic futures.

Physics-based character animation has seen significant advances in recent years with the adoption of Deep Reinforcement Learning (DRL). However, DRL-based learning methods are usually computationally expensive and their performance crucially depends on the choice of hyperparameters. Tuning hyperparameters for these methods often requires repetitive training of control policies, which is even more computationally prohibitive. In this work, we propose a novel Curriculum-based Multi-Fidelity Bayesian Optimization framework (CMFBO) for efficient hyperparameter optimization of DRL-based character control systems. Using curriculum-based task difficulty as fidelity criterion, our method improves searching efficiency by gradually pruning search space through evaluation on easier motor skill tasks. We evaluate our method on two physics-based character control tasks: character morphology optimization and hyperparameter tuning of DeepMimic. Our algorithm significantly outperforms state-of-the-art hyperparameter optimization methods applicable for physics-based character animation. In particular, we show that hyperparameters optimized through our algorithm result in at least 5x efficiency gain comparing to author-released settings in DeepMimic.

A wide range of modern science and engineering applications are formulated as optimization problems with a system of partial differential equations (PDEs) as constraints. These PDE-constrained optimization problems are typically solved in a standard discretize-then-optimize approach. In many industry applications that require high-resolution solutions, the discretized constraints can easily have millions or even billions of variables, making it very slow for the standard iterative optimizer to solve the exact gradients. In this work, we propose a general framework to speed up PDE-constrained optimization using online neural synthetic gradients (ONSG) with a novel two-scale optimization scheme. We successfully apply our ONSG framework to computational morphogenesis, a representative and challenging class of PDE-constrained optimization problems. Extensive experiments have demonstrated that our method can significantly speed up computational morphogenesis (also known as topology optimization), and meanwhile maintain the quality of final solution compared to the standard optimizer. On a large-scale 3D optimal design problem with around 1,400,000 design variables, our method achieves up to 7.5x speedup while producing optimized designs with comparable objectives.

Recently, our proposed recurrent neural network (RNN) based all deep learning minimum variance distortionless response (ADL-MVDR) beamformer method yielded superior performance over the conventional MVDR by replacing the matrix inversion and eigenvalue decomposition with two recurrent neural networks. In this work, we present a self-attentive RNN beamformer to further improve our previous RNN-based beamformer by leveraging on the powerful modeling capability of self-attention. Temporal-spatial self-attention module is proposed to better learn the beamforming weights from the speech and noise spatial covariance matrices. The temporal self-attention module could help RNN to learn global statistics of covariance matrices. The spatial self-attention module is designed to attend on the cross-channel correlation in the covariance matrices. Furthermore, a multi-channel input with multi-speaker directional features and multi-speaker speech separation outputs (MIMO) model is developed to improve the inference efficiency. The evaluations demonstrate that our proposed MIMO self-attentive RNN beamformer improves both the automatic speech recognition (ASR) accuracy and the perceptual estimation of speech quality (PESQ) against prior arts.

Transformer based architectures are recently used for the task of answering questions over tables. In order to improve the accuracy on this task, specialized pre-training techniques have been developed and applied on millions of open-domain web tables. In this paper, we propose two novel approaches demonstrating that one can achieve superior performance on table QA task without even using any of these specialized pre-training techniques. The first model, called RCI interaction, leverages a transformer based architecture that independently classifies rows and columns to identify relevant cells. While this model yields extremely high accuracy at finding cell values on recent benchmarks, a second model we propose, called RCI representation, provides a significant efficiency advantage for online QA systems over tables by materializing embeddings for existing tables. Experiments on recent benchmarks prove that the proposed methods can effectively locate cell values on tables (up to ~98% Hit@1 accuracy on WikiSQL lookup questions). Also, the interaction model outperforms the state-of-the-art transformer based approaches, pre-trained on very large table corpora (TAPAS and TaBERT), achieving ~3.4% and ~18.86% additional precision improvement on the standard WikiSQL benchmark.

Python is the most widely used programming language today. When it comes to solving data science tasks and challenges, Python never ceases to surprise its users. Most data scientists are already leveraging the power of Python programming every day. Python is an easy-to-learn, easy-to-debug, widely used, object-oriented, open-source, high-performance language, and there are many more benefits to Python programming. Python has been built with extraordinary Python libraries for data science that are used by programmers every day in solving problems.Here today, We have curated a list of best 15 Python libraries that helps in Data Science and its periphery, when to use them, their advantages and best tutorials to learn them. Pandas stand for Python Data Analysis Library. Pandas is the best tool for data wrangling or munging.

Please join us for a virtual Google workshop on "Conceptual Understanding of Deep Learning" When: May 17th 9am-4pm PST. Goal: How does the Brain/Mind (perhaps even an artificial one) work at an algorithmic level? While deep learning has produced tremendous technological strides in recent decades, there is an unsettling feeling of a lack of "conceptual" understanding of why it works and to what extent it will work in the current form. The goal of the workshop is to bring together theorists and practitioners to develop an understanding of the right algorithmic view of deep learning, characterizing the class of functions that can be learned, coming up with the right learning architecture that may (provably) learn multiple functions, concepts and remember them over time as humans do, theoretical understanding of language, logic, RL, meta learning and lifelong learning. The speakers and panelists include Turing award winners Geoffrey Hinton, Leslie Valiant, and Godel Prize winner Christos Papadimitriou (full-details).

I am an intermediate in ML and I know the theory of several supervised and unsupervised ML algorithms as well as deep learning. However, I lack huge amount of practical skills which is what I am working on right now. However, I am lost as to what to practice exactly. I am mostly interested in deep learning but I am seeing how essential it is to know how to implement other algorithms as well like random forests, SVM.. etc. Do I practice DL and other ML algorithms simultaneously (as mastering one area is almost impossible and is done across many years of experience) or should I first focus on one area (say computer vision) and then move on to the rest?

GPT-3 (Generative Pre-trained Transformer 3), is an artificial intelligence language generator that uses deep learning to produce human-like output. The high quality of its text is very difficult to distinguish from a human's. Many scientists, researchers and engineers (including Stephen Hawking and Elon Musk) have warned of A.I.'s potential dangers and called for steps to mitigate risk. And deep-learning critic Gary Marcus, has said that GPT-3's "comprehension of the world is often seriously off, which means you can never really trust what it says."

Waymo, Alphabet's self-driving car subsidiary, has reshuffled its top executive lineup. John Krafcik, Waymo's CEO since 2015, announced on April 2 that he would be stepping down from his role. Krafcik is being replaced by former COO Tekedra Mawakana and former CTO Dmitri Dolgov and will remain as an advisor to the company. "[With] the fully autonomous Waymo One ride-hailing service open to all in our launch area of Metro Phoenix, and with the fifth generation of the Waymo Driver being prepared for deployment in ride-hailing and goods delivery, it's a wonderful opportunity for me to pass the baton to Tekedra and Dmitri as Waymo's co-CEOs," Krafcik wrote on LinkedIn. The change in leadership could have significant implications for Waymo, which has seen many ups and downs as it develops its driverless car business.