ISCAS 2022

111 papers
The 2022 International Symposium on Circuits and Systems (ISCAS) featured 120 papers emphasizing the optimization of machine learning models for hardware implementation. Research themes focused on parameter-efficient fine-tuning, selective layer adaptation, and the sparsification of transformer bottlenecks. A major portion of the work addressed edge computing, specifically ultra-low-power inference via sub-byte quantization and resource-aware intelligence for autonomous systems. In the field of signal processing, research highlighted advancements in video coding, such as deep parameter adaptation and block partitioning techniques extending beyond Versatile Video Coding (VVC). Hardware developments included reconfigurable hybrid ADC architectures for computing-in-memory and heterogeneous processors for software-defined RF convergence. Additional studies explored no-reference quality assessment for super-resolution images and task-aware compression for health monitoring. Prolific authors contributing to these developments included Chenglin Li, Hongkai Xiong, Junni Zou, and Wenrui Dai.

Top Research Directions

Research directions most represented by ISCAS 2022 papers.

Energy Efficient Computing with Spiking Neural Networks 2
Recurrent Feedback for Associative Pattern Completion 1
Analytical Error Compensation for Approximate Deep Learning Hardware 1
The Sequential Bottleneck Shift in Accelerated Systems 1
Spatially Adaptive Bit Allocation for Image Compression 1
Multi Core Energy Scaling for Edge Inference 1
Sparse Neural Coding via Competitive Activation 1
Hardware Efficient Approximation of Non Linear Operations 1
Mitigating the Feed-Forward Network Bottleneck in Transformers 1
Robust Deep Learning on Noisy Analog Memory Crossbars 1
Emulating Synaptic Plasticity with Memristive Hardware 1
Unsupervised Feature Learning with Spike Timing Dependent Plasticity 1
Learnable Scaling Parameters for Neural Network Quantization 1
Scaling Model Evaluation Beyond Simple Classification Tasks 1
Conversion of Artificial Neural Networks to Spiking Networks 1

Top Authors

AuthorPapers
Li zhang4
Yibo Fan4
Bing Zeng3
Chenglin Li3
Chenglin Li3
Hongkai Xiong3
Jie Yang3
Li Du3
Shuyuan Zhu3
Wenrui Dai3
Yuan Du3
Chang Liu2
Dong Liu2
Hanli Wang2
Jason Eshraghian2
Jiaying Liu2
Jing Liu2
Li Song2
Mohamad Sawan2
Pengjie Tang2

Impactful Papers

Papers from ISCAS 2022 that shaped research clusters.

Foundational

Influential

Papers

Title
SpikingSIM: A Bio-Inspired Spiking Simulator
Stochastic dendrites enable online learning in mixed-signal neuromorphic processing systems
Structured and tiled-based pruning of Deep Learning models targeting FPGA implementations
TAFNet: A Three-Stream Adaptive Fusion Network for RGB-T Crowd Counting
The Hierarchical Ensemble Model for Network Intrusion Detection in the Real-world Dataset
Total Ionization Dose Measurement onboard a 1U CubeSat in Low Earth Orbit
Towards Task-aware Signal Compression for Efficient Continuous Health Monitoring
Two-Stage Energy Efficiency Optimization of Switched-Capacitor Converters for IoT Systems
ViSTRA3: Video Coding with Deep Parameter Adaptation and Post Processing
Wavelet-Based Learned Scalable Video Coding
Wearable Bowel Sound Monitoring with Quality Enhancement using U-net

« prev Page 3/3