Technical Program

Thursday, 21 October, 11:30AM - 1:00PM

T1-1: Efficiency

  • Leveraging Prediction Confidence For Versatile Optimizations to CNNs (R)
  • Federated matched averaging with information-gain based parameter sampling (R)
  • A Deep Learning framework for Interoperable Machine Learning (I)
  • Towards Accelerating Offline RL based Recommender Systems (I)
  • On Handling Class Imbalance in Continual Learning based Network Intrusion Detection Systems (R)
  • Generalized Weight Agnostic Neural Networks for Configurable and Continual Autonomous Systems (R)

Thursday, 21 October, 2:00PM - 3:30PM

T1-2-1: Socio-Economic Applications + demo

Session Chair: Steven Euijong Whang, KAIST, Korea

  • Stacked Sparse Autoencoder and Machine Learning Based Anxiety Classification Using EEG Signals (R)
  • InuAid Chest X-rays AI Product (I)
  • Real-time Social Distancing Detection System with Auto-calibration using Pose Information (Demo)
  • Measuring Viscosity of Fluids: A Deep Learning Approach Using a CNN-RNN Architecture (R)
  • Targeted Upskilling Framework based on Player Mistake Context in Online Skill Gaming Platforms (R)
  • Source-Code Similarity Measurement: Syntax Tree Fingerprinting for Automated Evaluation (R)

Thursday, 21 October, 3:45PM - 4:30PM

T1-2-2: ML and Security

Session Chair: Sparsh Mittal, IIT Roorkey, India

  • Inferring DNN layer-types through a hardware performance counters based side channel attack (R)
  • BOND: Efficient and Frugal DL Model Co-design for Botnet detection on IoT Gateways (R)
  • CryptInfer - Enabling Practical Inference on Encrypted Skin Lesion Images for Melanoma Detection (R)

Friday, 22 October, 10:00 AM - 11:30 AM

T1-3: NLP and Information Retrieval

Session Chair: Vinay Ribeiro, IIT Bombay, India

  • A Word Embedding Topic Model for Robust Inference of Topics and Visualization (R)
  • Word-level beam search decoding and correction algorithm (WLBS) for end-to-end ASR (R)
  • Few shot learning for cross-lingual isolated word recognition (R)
  • Class-Based Order-Independent Models of Natural Language for Bayesian Auto-Complete Inference (I)
  • A Hybrid Approach for Offline A/B Evaluation for Item Ranking Algorithms in Recommendation Systems (I)

Saturday, 23 October, 12:00 PM - 01:15 PM

T1-5: Systems for ML Opt

  • Bridging between Data Science and Performance Analysis: Tracing of Jupyter Notebooks (R)
  • Scarecrow - Intelligent Annotation platform for Engine Health Management (I)
  • Optimizing Convolutions for an Inference Accelerator [Case Study: Intel’s NNP-I 1000 DL Compute Grid] (I)
  • Resource Constrained Neural Networks for Direction-of-Arrival Estimation in Micro-controllers (R)
  • VeRLPy: Python Library for Verification of Digital Designs with Reinforcement Learning (R)