Systems and Architecture for ML

Recent Publications

Systems and Hardware Support

[ASPLOS 2024] Amanda: Unified Instrumentation Framework for Deep Neural Networks

[ISCA 2023] OliVe: Accelerating Large Language Models via Hardware-friendly Outlier-Victim Pair Quantization

[HPCA 2022] S2TA: Exploiting Structured Sparsity for Energy-Efficient Mobile CNN Acceleration

[MICRO 2022] ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization

[ICLR 2022] SQuant: On-the-Fly Data-Free Quantization via Diagonal Hessian Approximation

[IISWC 2021] Characterizing and Demystifying the Implicit Convolution Algorithm on Commercial Matrix-Multiplication Accelerators

[ISPASS 2020] A Systematic Methodology for Characterizing Scalability of DNN Accelerators using SCALE-Sim

[DAC 2020] Balancing Efficiency and Flexibility for DNN Acceleration via Temporal GPU-Systolic Array Integration

[SC 2020] Accelerating Sparse DNN Models Without Hardware-Support via Tile-wise Sparsity

[IPDPS 2018] BitFlow: Exploiting Vector Parallelism for Binary Neural Networks on CPU

Adversarial Robustness

[MICRO 2020] Ptolemy: Architecture Support for Robust Deep Learning

[CVPR 2019] Adversarial Defense Through Network Profiling Based Path Extraction

Resource-Guaranteeing Deep Learning

[CVPR 2020] Automatic Neural Network Compression by Sparsity-Quantization Joint Learning: A Constrained Optimization-based Approach

[CVPR 2019] ECC: Energy-Constrained Deep Neural Network Compression via a Bilinear Regression Model

[ICLR 2019] Energy-Constrained Compression for Deep Neural Networks via Weighted Sparse Projection and Layer Input Masking