A robot can now grasp an object more effectively than ever before, but once it has the object what happens next? We show that a mild relaxation of the task and workspace constraints implicit in existing object grasping datasets can cause neural network based grasping algorithms to fail on even a simple block stacking task when executed under more realistic circumstances. To address this, we introduce the JHU CoSTAR Block Stacking Dataset (BSD), where a robot interacts with 5.1 cm colored blocks to complete an order-fulfillment style block stacking task. It contains dynamic scenes and real time-series data in a less constrained environment than comparable datasets. There are nearly 12,000 stacking attempts and over 2 million frames of real data. We discuss the ways in which this dataset provides a valuable resource for a broad range of other topics of investigation. We find that hand-designed neural networks that work on prior datasets do not generalize to this task. Thus, to establish a baseline for this dataset, we demonstrate an automated search of neural network based models using a novel multiple-input HyperTree MetaModel, and find a final model which makes reasonable 3D pose predictions for grasping and stacking on our dataset. The CoSTAR BSD, code, and instructions are available at https://sites.google.com/site/costardataset.

Slimmable networks are a family of neural networks that can instantly adjust the runtime width. The width can be chosen from a predefined widths set to adaptively optimize accuracy-efficiency trade-offs at runtime. In this work, we propose a systematic approach to train universally slimmable networks (US-Nets), extending slimmable networks to execute at arbitrary width, and generalizing to networks both with and without batch normalization layers. We further propose two improved training techniques for US-Nets, named the sandwich rule and inplace distillation, to enhance training process and boost testing accuracy. We show improved performance of universally slimmable MobileNet v1 and MobileNet v2 on ImageNet classification task, compared with individually trained ones and 4-switch slimmable network baselines. We also evaluate the proposed US-Nets and improved training techniques on tasks of image super-resolution and deep reinforcement learning. Extensive ablation experiments on these representative tasks demonstrate the effectiveness of our proposed methods. Our discovery opens up the possibility to directly evaluate FLOPs-Accuracy spectrum of network architectures. Code and models will be available at: https://github.com/JiahuiYu/slimmable_networks

We present a Reinforcement Learning (RL) methodology to bypass Google reCAPTCHA v3. We formulate the problem as a grid world where the agent learns how to move the mouse and click on the reCAPTCHA button to receive a high score. We study the performance of the agent when we vary the cell size of the grid world and show that the performance drops when the agent takes big steps toward the goal. Finally, we used a divide and conquer strategy to defeat the reCAPTCHA system for any grid resolution. Our proposed method achieves a success rate of 97.4% on a 100x100 grid and 96.7% on a 1000x1000 screen resolution.

Large-scale distributed training of neural networks is often limited by network bandwidth, wherein the communication time overwhelms the local computation time. Motivated by the success of sketching methods in sub-linear/streaming algorithms, we propose a sketching-based approach to minimize the communication costs between nodes without losing accuracy. In our proposed method, workers in a distributed, synchronous training setting send sketches of their gradient vectors to the parameter server instead of the full gradient vector. Leveraging the theoretical properties of sketches, we show that this method recovers the favorable convergence guarantees of single-machine top-$k$ SGD. Furthermore, when applied to a model with $d$ dimensions on $W$ workers, our method requires only $\Theta(kW)$ bytes of communication, compared to $\Omega(dW)$ for vanilla distributed SGD. To validate our method, we run experiments using a residual network trained on the CIFAR-10 dataset. We achieve no drop in validation accuracy with a compression ratio of 4, or about 1 percentage point drop with a compression ratio of 8. We also demonstrate that our method scales to many workers.

Optimizing storage assignment is a central problem in warehousing. Past literature has shown the superiority of the Duration-of-Stay (DoS) method in assigning pallets, but the methodology requires perfect prior knowledge of DoS for each pallet, which is unknown and uncertain under realistic conditions. In this paper, we introduce a new framework that accounts for such uncertainty using a novel combination of convolutional and recurrent neural network models, ParallelNet. Through collaboration with a large cold storage company, we demonstrate ParallelNet achieves a 10% decrease in MSLE and 29% decrease in MAPE compared to CNN-LSTM on unseen future shipments, and suffers less performance decay as time increases. The framework is then integrated into a first-of-its-kind Storage Assignment system, which is being piloted in warehouses across the country.

This paper proposes a method to use deep neural networks as end-to-end open-set classifiers. It is based on intra-class data splitting. In open-set recognition, only samples from a limited number of known classes are available for training. During inference, an open-set classifier must reject samples from unknown classes while correctly classifying samples from known classes. The proposed method splits given data into typical and atypical normal subsets by using a closed-set classifier. This enables to model the abnormal classes by atypical normal samples. Accordingly, the open-set recognition problem is reformulated into a traditional classification problem. In addition, a closed-set regularization is proposed to guarantee a high closed-set classification performance. Intensive experiments on five well-known image datasets showed the effectiveness of the proposed method which outperformed the baselines and achieved a distinct improvement over the state-of-the-art methods.

In complex processes, various events can happen in different sequences. The prediction of the next event activity given an a-priori process state is of importance in such processes. Recent methods leverage deep learning techniques such as recurrent neural networks to predict event activities from raw process logs. However, deep learning techniques cannot efficiently model logical behaviors of complex processes. In this paper, we take advantage of Petri nets as a powerful tool in modeling logical behaviors of complex processes. We propose an approach which first discovers Petri nets from event logs utilizing a recent process mining algorithm. In a second step, we enhance the obtained model with time decay functions to create timed process state samples. Finally, we use these samples in combination with token movement counters and Petri net markings to train a deep learning model that predicts the next event activity. We demonstrate significant performance improvements and outperform the state-of-the-art methods on eight out of nine real-world benchmark event logs in accuracy.

OpenAI may not be quite so open going forward. The former nonprofit announced today that it is restructuring as a "capped-profit" company that cuts returns from investments past a certain point. But some worry that this move -- or rather the way they made it -- may result in making the innovative company no different from the other AI startups out there. From now on, profits from any investment in the OpenAI LP (limited partnership, not limited profit) will be passed on to an overarching nonprofit company, which will disperse them as it sees fit. Profits in excess of a 100x return, that is.

The evolution of artificial intelligence (AI) grew with the complexity of the languages available for development. In 1959, Arthur Samuel developed a self-learning checkers program at IBM on an IBM 701 computer using the native instructions of the machine (quite a feat given search trees and alpha-beta pruning). But today, AI is developed using various languages, from Lisp to Python to R. This article explores the languages that evolved for AI and machine learning. The programming languages that are used to build AI and machine learning applications vary. Each application has its own constraints and requirements, and some languages are better than others in particular problem domains.

Stories are the essence of human culture. They contain information about our past, and theories of the future. They allow us to delve into inner workings and subtleties of the human mind, discovering aspects of it that are impossible to analyze traditionally. While stories have always been written by our species in the past, with the growth and research in fields of deep learning, we see computer programs being able to write stories, and use language as humans do, but how? When you were reading the last two paragraphs, you based your understanding of every word off of the previous sentences and words.