Adversarial robustness corresponds to the susceptibility of deep neural networks to imperceptible perturbations made at test time. In the context of image tasks, many algorithms have been proposed to make neural networks robust to adversarial perturbations made to the input pixels. These perturbations are typically measured in an ℓpnorm. However, robustness often holds only for the specific attack used for training. In this work we extend the above setting to consider the problem of training of deep neural networks that can be made simultaneously robust to perturbations applied in multiple natural representations spaces. For the case of image data, examples include the standard pixel representation as well as the representation in the discrete cosine transform (DCT) basis. We design a theoretically sound algorithm with formal guarantees for the above problem. Furthermore, our guarantees also hold when the goal is to require robustness with respect to multiple ℓpnorm based attacks. We then derive an efficient practical implementation and demonstrate the effectiveness of our approach on standard datasets for image classification.1

Developing agile behaviors for legged robots re-mains a challenging problem. While deep reinforcement learning is a promising approach, learning truly agile behaviors typically requires tedious reward shaping and careful curriculum design. We formulate agile locomotion as a multi-stage learning problem in which a mentor guides the agent throughout the training. The mentor is optimized to place a checkpoint to guide the movement of the robot’s center of mass while the student (i.e. the robot) learns to reach these checkpoints. Once the student can solve the task, we teach the student to perform the task without the mentor. We evaluate our proposed learning system with a simulated quadruped robot on a course consisting of randomly generated gaps and hurdles. Our method significantly outperforms a single-stage RL baseline without a mentor, and the quadruped robot can agilely run and jump across gaps and obstacles. Finally, we present a detailed analysis of the learned behaviors’ feasibility and efficiency.

We present Neural Structured Learning (NSL) in TensorFlow [1], a new learning paradigm to train neural networks by leveraging structured signals in addition to feature inputs. Structure can be explicit as represented by a graph, or implicit, either induced by adversarial perturbation or inferred using techniques like embedding learning. NSL is open-sourced as part of the TensorFlow [2] ecosystem and is widely used in Google across many products and services. In this tutorial, we provide an overview of the NSL framework including various libraries, tools, and APIs as well as demonstrate the practical use of NSL in different applications. The NSL website is hosted at www.tensorflow.org/neural_structured_learning, which includes details about the theoretical foundations of the technology, extensive API documentation, and hands-on tutorials.

We present Neural Structured Learning (NSL) in TensorFlow [2], a new learning paradigm to train neural networks by leveraging structured signals in addition to feature inputs. Structure can be explicit as represented by a graph, or implicit, either induced by adversarial perturbation or inferred using techniques like embedding learning. NSL is open-sourced as part of the TensorFlow [3] ecosystem and is widely used in Google across many products and services. In this tutorial, we provide an overview of the NSL framework including various libraries, tools, and APIs as well as demonstrate the practical use of NSL in different applications. The NSL website is hosted at www.tensorflow.org/neural_structured_learning, which includes details about the theoretical foundations of the technology, extensive API documentation, and hands-on tutorials.

The widespread need to assess fall risk has generated need for a reliable, inexpensive, and quantifiable clinical measure. In this paper, we propose mStroke, a practical, accurate, and mobile health system that remotely measures fall risk estimation for post stroke patients. Our proposed system is delivered as an application (App) running on our hardware system consisting of two bluetooth low energy (BLE) modular sensor devices and a smartphone. This system uses accelerometer, gyroscope, and magnetometer to measure fall risk within the functional reach test, an accurate single-tasked clinical tool for fall risk assessment. The proposed system has been extensively tested using emulated and real data from Physical Therapy students as well as stroke patients.

We present an approach to wearable sensor-based assessment of fall risk for post stroke patients. The goal is to use inertial measurement unit to automate the fall risk estimation of the functional reach test. In this paper, we present the current stage of our system, including the proposed Event Scoring technique for detecting events such as shuffling and taking steps during the test. The system is also equipped with our proposed signal processing tools for measuring the patient's balance. The proposed system is delivered as an application (App) running on our hardware system consisting of two bluetooth low energy modular sensor devices and a smart phone. This system has been extensively tested using emulated and real data from Physical Therapy students as well as stroke patients.

As the cost of IEEE 802.11 hardware continues to fall, the appeal of inserting unauthorized wireless access into enterprise networks grows. These rogue access points (AP) expose the enterprise network to a barrage of security vulnerabilities in that they are typically connected to a network port behind the firewall. Most of the current approaches to detecting rogue AP are rudimentary and are easily evaded by hackers. We propose the use of temporal traffic characteristics to detect rogue AP at a central location. This detection is independent of the wireless technology (IEEE 802.11a, 802.11b, or 802.11g), is scalable, does not possess the inefficiencies of the current solutions, and is independent of the signal range of the rogue AP.

The authors present a method for segmenting speech waveforms containing several speakers into utterances, each from one individual, and then identifying each utterance as coming from a specific individual or group of individuals. The procedure is unsupervised in that there is no training set, and sequential in that information obtained in early stages of the process is utilized in later stages.>

Several attempts to improve recognition accuracy with the use of supervised clustering techniques are described. These techniques modify the distance metric and/or the clustering procedure in a discrete hidden Markov model recognition system in an attempt to improve phonetic modeling. Three techniques considered are linear discriminant analysis, a hierarchical supervised vector quantization technique, and Kohonen's LVQ2 technique. All experiments were performed on the DARPA resource management speech corpus using the BBN BYBLOS system. Even though the techniques improved the phonetic recognition capability of the vector quantization, the overall word and sentence recognition accuracy did not improve.>