Automated surgical workflow analysis and understanding can assist surgeons to standardize procedures and enhance post-surgical assessment and indexing, as well as, interventional monitoring. Computer-assisted interventional (CAI) systems based on video can perform workflow estimation through surgical instruments' recognition while linking them to an ontology of procedural phases. In this work, we adopt a deep learning paradigm to detect surgical instruments in cataract surgery videos which in turn feed a surgical phase inference recurrent network that encodes temporal aspects of phase steps within the phase classification. Our models present comparable to state-of-the-art results for surgical tool detection and phase recognition with accuracies of 99 and 78% respectively.

We propose to take a novel approach to robot system design where each building block of a larger system is represented as a differentiable program, i.e. a deep neural network. This representation allows for integrating algorithmic planning and deep learning in a principled manner, and thus combine the benefits of model-free and model-based methods. We apply the proposed approach to a challenging partially observable robot navigation task. The robot must navigate to a goal in a previously unseen 3-D environment without knowing its initial location, and instead relying on a 2-D floor map and visual observations from an onboard camera. We introduce the Navigation Networks (NavNets) that encode state estimation, planning and acting in a single, end-to-end trainable recurrent neural network. In preliminary simulation experiments we successfully trained navigation networks to solve the challenging partially observable navigation task.

In this paper we show that restricting the representation-layer of a Recurrent Neural Network (RNN) improves accuracy and reduces the depth of recursive training procedures in partially observable domains. Artificial Neural Networks have been shown to learn useful state representations for high-dimensional visual and continuous control domains. If the the tasks at hand exhibits long depends back in time, these instantaneous feed-forward approaches are augmented with recurrent connections and trained with Back-prop Through Time (BPTT). This unrolled training can become computationally prohibitive if the dependency structure is long, and while recent work on LSTMs and GRUs has improved upon naive training strategies, there is still room for improvements in computational efficiency and parameter sensitivity. In this paper we explore a simple modification to the classic RNN structure: restricting the state to be comprised of multi-step General Value Function predictions. We formulate an architecture called General Value Function Networks (GVFNs), and corresponding objective that generalizes beyond previous approaches. We show that our GVFNs are significantly more robust to train, and facilitate accurate prediction with no gradients needed back-in-time in domains with substantial long-term dependences.

The user uses the control panel (A) to specify the detail information to visualize (e.g., level of detail, rule filters). The rule-based explanatory representation is visualized as a matrix (B), where each row represents a rule, and each column is a feature used in the rules. The user can also filter the data or use a customized input in the data filter (C) and navigate the filtered dataset in the data table (D). Abstract--With the growing adoption of machine learning techniques, there is a surge of research interest towards making machine learning systems more transparent and interpretable. Various visualizations have been developed to help model developers understand, diagnose, and refine machine learning models. However, a large number of potential but neglected users are the domain experts with little knowledge of machine learning but are expected to work with machine learning systems. In this paper, we present an interactive visualization technique to help users with little ...

The detection of software vulnerabilities (or vulnerabilities for short) is an important problem that has yet to be tackled, as manifested by many vulnerabilities reported on a daily basis. This calls for machine learning methods to automate vulnerability detection. Deep learning is attractive for this purpose because it does not require human experts to manually define features. Despite the tremendous success of deep learning in other domains, its applicability to vulnerability detection is not systematically understood. In order to fill this void, we propose the first systematic framework for using deep learning to detect vulnerabilities. The framework, dubbed Syntax-based, Semantics-based, and Vector Representations (SySeVR), focuses on obtaining program representations that can accommodate syntax and semantic information pertinent to vulnerabilities. Our experiments with 4 software products demonstrate the usefulness of the framework: we detect 15 vulnerabilities that are not reported in the National Vulnerability Database. Among these 15 vulnerabilities, 7 are unknown and have been reported to the vendors, and the other 8 have been "silently" patched by the vendors when releasing newer versions of the products.

We found an easy and quick post-learning method named "Icing on the Cake" to enhance a classification performance in deep learning. The method is that we train only the final classifier again after an ordinary training is done.

We generate pseudo-features by multivariate probability distributions obtained from feature maps in a low layer of trained deep neural networks. Then, we virtually augment the data of minor classes by the pseudo-features in order to overcome imbalanced data problems.

Most people find it a pain to receive parcels in wide time slots like, say, between 8 a.m. and 5 p.m., so when delivery startup Paack offered a service in which everyone could narrow that window down to one hour, with no extra charge, it had a challenge on its hands. The startup's routing engine worked but needed to be more efficient. Enter Prowler.io, a Cambridge, U.K.-based machine-learning startup that bills itself as a decision-making platform for any company with complex problems to solve. Paack's investors at Balderton in London introduced it to Prowler in February 2018, and within months its delivery vans and trucks were being coordinated by an intelligent, digital simulation. With the beta test over, Paack's CEO, Fernando Benito, sees a potential benefit to his bottom line. Some of his startup's deliveries are now 15% more efficient, he tells Forbes.

Recent advances in deep learning and exponential growth in the use of machine learning across application domains have made AI acceleration critically important. IBM Research has been building a pipeline of AI hardware accelerators to meet this need. At the 2018 VLSI Circuits Symposium, we presented a multi-TeraOPS accelerator core building block that can be scaled across a broad range of AI hardware systems. This digital AI core features a parallel architecture that ensures very high utilization and efficient compute engines that carefully leverage reduced precision. Approximate computing is a central tenet of our approach to harnessing "the physics of AI", in which highly energy-efficient computing gains are achieved by purpose-built architectures, initially using digital computations and later including analog and in-memory computing.

Energy Saving: In 2014, Google acquired an AI startup, DeepMind, to slash costs and improve efficiencies in its data centers. The AI engine automatically managed power usage by discovering and reporting inefficiencies across 120 data center variables - fans, cooling systems, windows etc. The results have been positive. Google has been able to reduce its total data center power consumption by 15 percent, saving the company millions over the next several years. Additionally, the company saved 40 percent alone on power consumed for cooling purposes.