• Publications
  • Influence
Gradient-Based Inference for Networks with Output Constraints
TLDR
In this paper, we present an inference method for neural networks that enforces deterministic constraints on outputs without performing rule-based post-processing or expensive discrete search. Expand
  • 16
  • 3
  • PDF
Towards Semi-Supervised Learning for Deep Semantic Role Labeling
TLDR
We propose a semi-supervised semantic role labeling method that outperforms the state-of-the-art in limited SRL training corpora. Expand
  • 10
  • 1
  • PDF
Correlating night-time satellite images with poverty and other census data of India and estimating future trends
TLDR
This paper proposes a method to correlate light intensity from images with state-wise poverty, population, GDP, and forest cover, and forecast future values of the same for each state. Expand
  • 4
  • 1
An LSTM Based System for Prediction of Human Activities with Durations
TLDR
We propose a deep learning system for predicting human activities and their durations in real-time on mobile devices. Expand
  • 14
Preventing Inadvertent Information Disclosures via Automatic Security Policies
TLDR
We propose an algorithm that analyzes the sensitive information and historic access permissions to identify content-access correspondence via a novel multi-label classifier formulation that is capable of recommending policies/access permissions for any new document. Expand
  • 3
Learning Rhyming Constraints using Structured Adversaries
TLDR
We propose an alternate approach that uses a structured discriminator to learn a poetry generator that directly captures rhyming constraints in a generative adversarial setup. Expand
  • 2
  • PDF
Improving Marketing Interactions by Mining Sequences
TLDR
We propose an approach based on sequence mining to identify marketing touch sequences that are most likely to lead to a stated marketing goal. Expand
  • 2
Modeling End-of-Online-Session From Streaming Data
TLDR
We address the problem of retargeting by using automated predictive models. Expand