Abstract

In this paper, we design and evaluate a convolutional autoencoder that perturbs an input face image to impart privacy to a subject. Specifically, the proposed autoencoder transforms an input face image such that the transformed image can be successfully used for face recognition but not for gender classification. In order to train this autoencoder, we propose a novel training scheme, referred to as semiadversarial training in this work. The training is facilitated by attaching a semi-adversarial module consisting of an auxiliary gender classifier and an auxiliary face matcher to the autoencoder. The objective function utilized for training this network has three terms: one to ensure that the perturbed image is a realistic face image; another to ensure that the gender attributes of the face are confounded; and a third to ensure that biometric recognition performance due to the perturbed image is not impacted. Extensive experiments confirm the efficacy of the proposed architecture in extending gender privacy to face images.

Abstract

The HITL paradigm has been extended as SITL (SocietyIn-The-Loop) to account for the broader role of AI in the society and vice versa. To open up these otherwise opaque systems and their nexus of interactions with humans, there is a need to make tools to program and debug the algorithmic social contract, a pact between various human stakeholders, mediated by machines. In this paper, we propose one such tool, the ESO-5W1H framework to adjudge the role of humans and machines in their respective interactions and to structure the underlying decision making process such that accountability and liability for each system action-interactions can be brought to the fore. We discuss the working of this conceptual framework in the context of three use cases: the Self-driving car, an AI-based jury, and Neural Networks

Abstract

An effective news recommendation system should harness the his- torical information of the user based on her interactions as well as the content of the articles. In this paper we propose a novel deep learning model for news recommendation which utilizes the content of the news articles as well as the sequence in which the articles were read by the user. To model both of these information, which are essentially of different types, we propose a simple yet effective architecture which utilizes a 3-dimensional Convolutional Neural Network which takes the word embeddings of the articles present in the user history as its input. Using such a method en- dows the model with the capability to automatically learn spatial (features of a particular article) as well as temporal features (fea- tures across articles read by a user) which signify the interest of the user. At test time, we use this in combination with a 2-dimensional Convolutional Neural Network for recommending articles to users. On a real-world dataset our method outperformed strong baselines which also model the news recommendation problem using neural networks.

Abstract

Science journalism is the art of conveying a detailed scientific research paper in a form that nonscientists can understand and appreciate while ensuring that its underlying information is conveyed accurately. It aims to transform jargon-laden scientific articles into a form that a common reader can comprehend while ensuring that the meaning of the article is retained. It plays a crucial role in making scientific content suitable for consumption by the public at large. Recent advances in Deep Learning research and it’s applications in natural language processing have made way to impressive results in Natural Language Generation. We leverage these advances to explore the possibility of their use in journalism, science journalism in particular, as comprehension of scientific content is much harder challenge than most of the other forms of content, like shorthand, which journalists use while writing articles. In this work, we introduce the problem of automated science journalism and present ways to automate some parts of the workflow by automatically generating the ‘title’ of a blog version of a scientific paper. We have built a corpus of 87, 328 pairs of research papers and their corresponding blogs from two science news aggregators and have used it to build ScienceBlogger - a pipeline-based architecture consisting of a two-stage mechanism to generate the blog titles. To demonstrate the models, we built an interactive tool, where a user can give abstract and title of a research paper, which would be processed by our APIs to produce a blog title, along with some relevant information about the model used for the generation. Evaluation using standard metrics indicate viability of the proposed system.

Abstract

With news stories coming from a variety of sources, it is crucial for news aggregators to present interesting articles to the user to maximize their engagement. This creates the need for a news recommendation system which understands the content of the articles as well as accounts for the users’ preferences. Methods such as Collaborative Filtering, which are well known for general recommendations, are not suitable for news because of the short life span of articles and because of the large number of articles published each day. Apart from this, such methods do not harness the information present in the sequence in which the articles are read by the user and hence are unable to account for the specific and generic interests of the user which may keep changing with time. In order to address these issues for news recommendation, we propose the Recurrent Attentive Recommendation Engine (RARE). RARE consists of two components and utilizes the distributed representations of news articles. The first component is used to model the user’s sequential behaviour of news reading in order to understand her general interests, i.e., to get a summary of her interests. The second component utilizes an article level attention mechanism to understand her specific interests. We feed the information obtained from both the components to a Siamese Network in order to make predictions which pertain to

Abstract

The Virtual Labs project (vlabs) is a country wide effort to provide access to internet and browser based virtual laboratories to science and engineering college students in India. There are currently over 800 experiments packaged in 70 vlabs in more than ten science and engineering disciplines. More than five hundred workshops on these virtual labs have been conducted and the experiments have been used over 2 million times. Vlabs are an important Government initiative in higher learning in engineering in India. This paper examines virtual labs from a usability perspective, focusing on vlabs. Usability of user-interface elements (technical usability) as well as how they aid learning (pedagogical usability) has been considered. We adapt Nielsen's technical usability heuristics and Nokelainen's pedagogical usability criteria to virtual labs and evaluate the usability of the top three most used vlabs. We construct checklists for each of these criteria and evaluate the vlabs against this checklist. Our analysis reveals a lack of usability focus in lab design. Lab design and implementation guidelines seem to be deficient in covering key usability criteria. We make suggestions on how to quickly improve the usability of these labs.

Abstract

In this paper, we investigate the sum-rate maximization problem in a cooperative relay network, which employs non-orthogonal multiple access (NOMA). Specifically, we address the problem of optimal power allocation and relay precoder design. The corresponding optimization problem turns out to be non-convex. In order to get around the difficulty to deal with non-convexity, we propose a minorization-maximization technique based algorithm to jointly optimize source power allocation and relay precoder design. We then solve the problem to optimally allocate power at source and design the precoder at relay. For the case of two destination nodes, we propose a grid search based approach. Simulation results are provided to illustrate the performance of the proposed algorithm.

Abstract

Two dimensional (2-D) direction of arrival (DOA) estimation is a fundamental problem in array signal processing with wide range of applications. In this paper, an approach to 2-D DOA estimation with a modified parallel coprime linear subarrays using MUSIC algorithm and least squares is proposed. A virtual difference co-array is synthesized by vectorizing the crosscovariance matrix of sub-array data. In the proposed method, the 2-D DOA estimation problem is decomposed as two one dimensional (1-D) DOA estimation problems in which azimuth and elevation DOAs are estimated and automatically paired using MUSIC and Least Squares (LS). Simulation results are presented to verify the performance of the proposed method and the improvements resulting from the proposed array geometry.

Abstract

Monocular SLAM refers to using a single camera to estimate robot ego motion while building a map of the environment. While Monocular SLAM is a well studied problem, automating Monocular SLAM by integrating it with trajectory planning frameworks is particularly challenging. This paper presents a novel formulation based on Reinforcement Learning (RL) that generates fail safe trajectories wherein the SLAM generated outputs do not deviate largely from their true values. Quintessentially, the RL framework successfully learns the otherwise complex relation between perceptual inputs and motor actions and uses this knowledge to generate trajectories that do not cause failure of SLAM. We show systematically in simulations how the quality of the SLAM dramatically improves when trajectories are computed using RL. Our method scales effectively across Monocular SLAM frameworks in both simulation and in real world experiments with a mobile robot.

Abstract

We propose a novel motion segmentation formulation over spatio-temporal depth images obtained from stereo sequences that segments multiple motion models in the scene in an unsupervised manner . The motion segmentation is obtained at frame rates that compete with the speed of the stereo depth computation. This is possible due to a decoupling framework that first delineates spatial clusters and subsequently assigns motion labels to each of these cluster with analysis of a novel motion graph model. A principled computation of the weights of the motion graph that signifies the relative shear and stretch between possible clusters lends itself to a high fidelity segmentation of the motion models in the scene. Keywords: Motion