In the 1970s, faced with the USSR's overwhelming superiority in numbers, the Department of Defense decided to compensate by focusing on high technology platforms. This led to the highly successful F-15, F-16, F-18, Abrams tanks, and Bradley fighting vehicles. Since then, the United States has continued to pursue cutting edge technology that has resulted in the highly capable F-22 and, when the testing and software development is complete, perhaps a highly capable F-35. Unfortunately, cost has accelerated faster than capabilities. And thus numbers have declined precipitously. The U.S. Air Force initially planned to buy 750 F-22s, but the high cost led Secretary of Defense Robert Gates to cap the program at 187. Nor has the Air Force been alone in pursuing top end systems. The Navy attempted an entirely new concept with "Streetfighter." Meant to be a low-cost, highly capable ship to replace the Navy's frigates and minesweepers for operations in brown water, it evolved into the Littoral Combat Ship.

Google has caved to pressure from its staff and abandoned a new AI ethics panel after hundreds demanded conservative members of the board were sacked for their views. The search giant announced last week that it was setting up a new board to tackle moral issues surrounding its use of the technology. It hoped to avoid controversies by using a broad spectrum of expertise to inform its future decisions, but the move has ironically stirred up a debacle of its own. Eight experts from outside the company were recruited and employees at the traditionally liberal leaning firm took issue with two of the appointees. More than 1,000 of its protest-prone workers signed an open letter objecting to specific board members, who they say are'anti-trans' and pro-military drones.

We introduce Bayesian least-squares policy iteration (BLSPI), an off-policy, model-free, policy iteration algorithm that uses the Bayesian least-squares temporal-difference (BLSTD) learning algorithm to evaluate policies. An online variant of BLSPI has been also proposed, called randomised BLSPI (RBLSPI), that improves its policy based on an incomplete policy evaluation step. In online setting, the exploration-exploitation dilemma should be addressed as we try to discover the optimal policy by using samples collected by ourselves. RBLSPI exploits the advantage of BLSTD to quantify our uncertainty about the value function. Inspired by Thompson sampling, RBLSPI first samples a value function from a posterior distribution over value functions, and then selects actions based on the sampled value function. The effectiveness and the exploration abilities of RBLSPI are demonstrated experimentally in several environments.

The regulatory process of Drosophila has been thoroughly studied for understanding a great variety of systems biology principles. While pattern-forming gene networks are further analysed in the transcription step, post-transcriptional events (e.g. translation, protein processing) play an important role in establishing protein expression patterns and levels. Since post-transcriptional regulation of gap genes in Drosophila depends on spatiotemporal interactions between mRNAs and gap proteins, proper physically-inspired stochastic models are required to study the existing link between both quantities. Previous research attempts have shown that the use of Gaussian processes (GPs) and differential equations leads to promising predictions when analysing regulatory networks. Here we aim at further investigating two types of physically-inspired GP models based on a reaction-diffusion equation where the main difference lies on whether the GP prior is placed. While one of them has been studied previously using gap protein data only, the other is novel and yields a simplistic approach requiring only the differentiation of kernel functions. In contrast to other stochastic frameworks, discretising the spatial space is not required here. Both GP models are tested under different conditions depending on the availability of gap gene mRNA expression data. Finally, their performances are assessed on a high-resolution dataset describing the blastoderm stage of the early embryo of Drosophila melanogaster.

This paper develops methodology for 3D radial visualization of high-dimensional datasets. Our display engine is called RadViz3D and extends the classic RadViz that visualizes multivariate data in the 2D plane by mapping every record to a point inside the unit circle. The classic RadViz display has equally-spaced anchor points on the unit circle, with each of them associated with an attribute or feature of the dataset. RadViz3D obtains equi-spaced anchor points exactly for the five Platonic solids and approximately for the other cases via a Fibonacci grid. We show that distributing anchor points at least approximately uniformly on the 3D unit sphere provides a better visualization than in 2D. We also propose a Max-Ratio Projection (MRP) method that utilizes the group information in high dimensions to provide distinctive lower-dimensional projections that are then displayed using Radviz3D. Our methodology is extended to datasets with discrete and mixed features where a generalized distributional transform is used in conjuction with copula models before applying MRP and RadViz3D visualization.

Data-dependent hashing has recently attracted attention due to being able to support efficient retrieval and storage of high-dimensional data such as documents, images, and videos. In this paper, we propose a novel learning-based hashing method called "Supervised Discrete Hashing with Relaxation" (SDHR) based on "Supervised Discrete Hashing" (SDH). SDH uses ordinary least squares regression and traditional zero-one matrix encoding of class label information as the regression target (code words), thus fixing the regression target. In SDHR, the regression target is instead optimized. The optimized regression target matrix satisfies a large margin constraint for correct classification of each example. Compared with SDH, which uses the traditional zero-one matrix, SDHR utilizes the learned regression target matrix and, therefore, more accurately measures the classification error of the regression model and is more flexible. As expected, SDHR generally outperforms SDH. Experimental results on two large-scale image datasets (CIFAR-10 and MNIST) and a large-scale and challenging face dataset (FRGC) demonstrate the effectiveness and efficiency of SDHR.

Reinforcement learning algorithms are gaining popularity in fields where optimal scheduling is important, and oncology is not an exception. The complex and uncertain dynamics of cancer limit the performance of traditional model-based scheduling strategies like Optimal Control. Motivated by the recent success of model-free Deep Reinforcement Learning (DRL) in challenging control tasks and in medical treatments, we use Deep Q-Network (DQN) and Deep Deterministic Policy Gradient (DDPG) to design a personalized cancer chemotherapy schedule. We show that both of them succeed in the task and outperform the Optimal Control solution in the presence of uncertainty. Furthermore, we show that DDPG can exterminate cancer more efficiently than DQN due to its continuous action space. Finally, we provide some intuition regarding the amount of samples required for the training.

Imitation learning is an effective alternative approach to learn a policy when the reward function is sparse. In this paper, we consider a challenging setting where an agent and an expert use different actions from each other. We assume that the agent has access to a sparse reward function and state-only expert observations. We propose a method which gradually balances between the imitation learning cost and the reinforcement learning objective. In addition, this method adapts the agent's policy based on either mimicking expert behavior or maximizing sparse reward. We show, through navigation scenarios, that (i) an agent is able to efficiently leverage sparse rewards to outperform standard state-only imitation learning, (ii) it can learn a policy even when its actions are different from the expert, and (iii) the performance of the agent is not bounded by that of the expert, due to the optimized usage of sparse rewards.

To help bridge the gap between internet vision-style problems and the goal of vision for embodied perception we instantiate a large-scale navigation task - Embodied Question Answering [1] in photo-realistic environments (Matterport 3D). We thoroughly study navigation policies that utilize 3D point clouds, RGB images, or their combination. Our analysis of these models reveals several key findings. We find that two seemingly naive navigation baselines, forward-only and random, are strong navigators and challenging to outperform, due to the specific choice of the evaluation setting presented by [1]. We find a novel lossweighting Figure 1: We extend EmbodiedQA [1] to photorealstic environments, scheme we call Inflection Weighting to be important our agent is spawned in a perceptually and semantically when training recurrent models for navigation with behavior novel environment and tasked with answering a cloning and are able to out perform the baselines question about that environment. We examine the agent's with this technique. We find that point clouds provide a ability to navigate the environment and answer the question richer signal than RGB images for learning obstacle avoidance, by perceiving its environment through point clouds, RGB motivating the use (and continued study) of 3D deep images, or a combination of the two.

Data-to-text generation can be conceptually divided into two parts: ordering and structuring the information (planning), and generating fluent language describing the information (realization). Modern neural generation systems conflate these two steps into a single end-to-end differentiable system. We propose to split the generation process into a symbolic text-planning stage that is faithful to the input, followed by a neural generation stage that focuses only on realization. For training a plan-to-text generator, we present a method for matching reference texts to their corresponding text plans. For inference time, we describe a method for selecting high-quality text plans for new inputs. We implement and evaluate our approach on the WebNLG benchmark. Our results demonstrate that decoupling text planning from neural realization indeed improves the system's reliability and adequacy while maintaining fluent output. We observe improvements both in BLEU scores and in manual evaluations. Another benefit of our approach is the ability to output diverse realizations of the same input, paving the way to explicit control over the generated text structure.