While accurate disease prediction from retinal fundus images is critical, collecting large amounts of high quality labeled training data to build such supervised models is difficult. Deep learning classifiers have led to high accuracy results across a wide variety of medical imaging problems, but they need large amounts of labeled data. Given a fundus image, we aim to evaluate various solutions for learning deep neural classifiers using small labeled data for three tasks related to eye disease prediction: (T1) predicting one of the five broad categories - diabetic retinopathy, age-related macular degeneration, glaucoma, melanoma and normal, (T2) predicting one of the 320 fine-grained disease sub-categories, (T3) generating a textual diagnosis. The problem is challenging because of small data size, need for predictions across multiple tasks, handling image variations, and large number of hyper-parameter choices. Modeling the problem under a multi-task learning (MTL) setup, we investigate the contributions of each of the proposed tasks while dealing with a small amount of labeled data. Further, we suggest a novel MTL-based teacher ensemble method for knowledge distillation. On a dataset of 7212 labeled and 35854 unlabeled images across 3502 patients, our technique obtains ~83% accuracy, ~75% top-5 accuracy and ~48 BLEU for tasks T1, T2and T3respectively. Even with 15% training data, our method outperforms baselines by 8.1, 3.2 and 11.2 points for the three tasks respectively.

User expectations of web search are changing. They are expecting search engines to answer questions, to be more conversational, and to offer means to complete tasks on their behalf. At the same time, to increase the breadth of tasks that personal digital assistants (PDAs), such as Microsoft»s Cortana or Amazon»s Alexa, are capable of, PDAs need to better utilize information about the world, a significant amount of which is available in the knowledge bases and answers built for search engines. It thus seems likely that the underlying systems that power web search and PDAs will converge. This demonstration presents a system that merges elements of traditional multi-turn dialog systems with web based question answering. This demo focuses on the automatic composition of semantic functional units, Botlets, to generate responses to user»s natural language (NL) queries. We show that such a system can be trained to combine information from search engine answers with PDA tasks to enable new user experiences.

Personal digital assistants are designed to assist users in easy information retrieval or execute the tasks they are interested in. The conversational medium implies an additional level of intelligence but typically these systems do not support any reference to the user's past interactions. We propose a domain-agnostic approach that enables the system to address queries referring to the past by using an information retrieval approach to rank various entities for a given query. We also add semantic enrichment to the recall process by augmenting the entities with information from a knowledge graph and leverage that in the retrieval process. We mined user interactions for the Cortana digital assistant to extract queries with location and business entities and show that our technique can achieve an accuracy of 89.8% for such recall queries.

Detecting impending failure of hard disks is an important prediction task which might help computer systems to prevent loss of data and performance degradation. Currently most of the hard drive vendors support self-monitoring, analysis and reporting technology (SMART) which are often considered unreliable for such tasks. The problem of finding alternatives to SMART for predicting disk failure is an area of active research. In this paper, we consider events recorded from live disks and show that it is possible to construct decision support systems which can detect such failures. It is desired that any such prediction methodology should have high accuracy and ease of interpretability. Black box models can deliver highly accurate solutions but do not provide an understanding of events which explains the decision given by it. To this end we explore rule based classifiers for predicting hard disk failures from various disk events. We show that it is possible to learn easy to understand rules, from disk events, which have extremely low false alarm rates on real world data.