Solving inverse problems with iterative algorithms is popular, especially for large data. Due to time constraints, the number of possible iterations is usually limited, potentially affecting the achievable accuracy. Given an error one is willing to tolerate, an important question is whether it is possible to modify the original iterations to obtain faster convergence to a minimizer achieving the allowed error without increasing the computational cost of each iteration considerably. Relying on recent recovery techniques developed for settings in which the desired signal belongs to some low-dimensional set, we show that using a coarse estimate of this set may lead to faster convergence at the cost of an additional reconstruction error related to the accuracy of the set approximation. Our theory ties to recent advances in sparse recovery, compressed sensing, and deep learning. Particularly, it may provide a possible explanation to the successful approximation of the l1-minimization solution by neural networks with layers representing iterations, as practiced in the learned iterative shrinkage-thresholding algorithm (LISTA).

Deep models that are both effective and explainable are desirable in many settings; prior explainable models have been unimodal, offering either image-based visualization of attention weights or text-based generation of post-hoc justifications. We propose a multimodal approach to explanation, and argue that the two modalities provide complementary explanatory strengths. We collect two new datasets to define and evaluate this task, and propose a novel model which can provide joint textual rationale generation and attention visualization. Our datasets define visual and textual justifications of a classification decision for activity recognition tasks (ACT-X) and for visual question answering tasks (VQA-X). We quantitatively show that training with the textual explanations not only yields better textual justification models, but also better localizes the evidence that supports the decision. We also qualitatively show cases where visual explanation is more insightful than textual explanation, and vice versa, supporting our thesis that multimodal explanation models offer significant benefits over unimodal approaches.

We identify obfuscated gradients, a kind of gradient masking, as a phenomenon that leads to a false sense of security in defenses against adversarial examples. While defenses that cause obfuscated gradients appear to defeat iterative optimization-based attacks, we find defenses relying on this effect can be circumvented. For each of the three types of obfuscated gradients we discover, we describe characteristic behaviors of defenses exhibiting this effect and develop attack techniques to overcome it. In a case study, examining non-certified white-box-secure defenses at ICLR 2018, we find obfuscated gradients are a common occurrence, with 7 of 8 defenses relying on obfuscated gradients. Our new attacks successfully circumvent 6 completely and 1 partially.

ABSTRACT Training a conventional automatic speech recognition (ASR) system to support multiple languages is challenging because the sub-word unit, lexicon and word inventories are typically language specific. In contrast, sequence-to-sequence models are well suited for multilingual ASR because they encapsulate an acoustic, pronunciation and language model jointly in a single network. In this work we present a single sequence-to-sequence ASR model trained on 9 different Indian languages, which have very little overlap in their scripts. Specifically, we take a union of language-specific grapheme sets and train a grapheme-based sequence-to-sequence model jointly on data from all languages. We find that this model, which is not explicitly given any information about language identity, improves recognition performance by 21% relative compared to analogous sequence-to-sequence models trained on each language individually. By modifying the model to accept a language identifier as an additional input feature, we further improve performance by an additional 7% relative and eliminate confusion between different languages.

The upcoming industrial revolution is anticipated to improve productivity, cost-efficiency and corporate competitiveness, made possible by connected sensors, IoT applications and the soon-to- released 5G networks to enable real time data connections from people, objects and contexts. Artificial intelligence (A.I) plays a critical role to collect and analyze massive volumes of digitized data to enable AI computing in visual inspection for smart factory, physical security for city safety, mobile edge computing for self-driving transits and adaptive streaming for video broadcasting. As the leading network solution provider, Lanner has taken a major step in network transformation by empowering next-generation SDN/NFV infrastructure with vCPE and NFVi platform. The next step of our mission is to take on the challenge of deep learning at the edge network with our advanced network platforms designed for high computing, high throughputs and high availability in A.I computing.

Yesterday I attended Spotify's Big Data & Machine Learning panel at Galvanize in NYC. Going in I had a few select questions, mainly why my #1 song on my 2016 Year in Review was "Venus" by Lady Gaga, despite all the times I set it to Private Mode for my ritualistic guilty pleasure listening sessions. More importantly than that, I came to learn how big data and machine learning are used by the engineers over at Spotify. With over 100M registered users worldwide, Spotify has many unique data and AI challenges. There's 25 billion data points within Spotify's database, 100 million music preferences to catalog and cater to, over 30M songs, and 2 billion playlists across 60 countries.

Artificial intelligence(AI) and machine learning(ML) are growing in popularity in the technology ecosystem. Just a few months ago, it was relatively simple to keep up with the developments in the AI and ML markets. Today, that seems like a daunting task for most technologists as the space have been evolving incredibly fast. The explosion of AI and ML platforms have created a very crowded market in which is very hard to distinguish signal from noise. However, despite the large number of AI technologies and startups, we can identify a few main categories that provide a good taxonomy to better understand the AI-ML markets.

Enhancer sequences regulate the expression of genes from afar by providing a binding platform for transcription factors, often in a tissue-specific or context-specific manner. Despite their importance in health and disease, our understanding of these DNA sequences, and their regulatory grammar, is limited. This impairs our ability to identify new enhancers along the genome, or to understand the effect of enhancer mutations and their role in genetic diseases. We trained deep Convolutional Neural Networks (CNN) to identify enhancer sequences in multiple species. We used multiple biological datasets, including simulated sequences, in vivo binding data of single transcription factors and genome-wide chromatin maps of active enhancers in 17 mammalian species.

Information contained on this page is provided by an independent third-party content provider. If you are affiliated with this page and would like it removed please contact pressreleases@franklyinc.com Loveland Innovations, maker of advanced data analytics solutions and drone-based data gathering tools announced the launch of the beta version of IMGING Detect, a deep learning engine built specifically for drone-based inspections. Deep learning is an advanced approach to artificial intelligence (A.I.) that allows IMGING to "learn" as it gathers more data, which makes IMGING more sophisticated and accurate each time it's used. This has vast implications for everything from damage detection to object and materials detection and beyond. The proprietary damage detection algorithms in IMGING are the most advanced currently available to the drone-based roof, building and property inspection space.

Abstract: The paper proposes and demonstrates a Deep Convolutional Neural Network (DCNN) architecture to identify users with disguised face attempting a fraudulent ATM transaction. The recent introduction of Disguised Face Identification (DFI) framework proves the applicability of deep neural networks for this very problem. All the ATMs nowadays incorporate a hidden camera in them and capture the footage of their users. However, it is impossible for the police to track down the impersonators with disguised faces from the ATM footage. The proposed deep convolutional neural network is trained to identify, in real time, whether the user in the captured image is trying to cloak his identity or not. The output of the DCNN is then reported to the ATM to take appropriate steps and prevent the swindler from completing the transaction.