Selected work

Representative Papers

DSLR-CNN: Efficient CNN Acceleration Using Digit-Serial Left-to-Right Arithmetic

Jan 03, 2025IEEE Access

To address energy-efficiency and latency bottlenecks in CNN hardware acceleration, this paper proposes DSLRCNN, a domain-specific accelerator leveraging left-to-right (LR) digit-serial arithmetic. It innovatively integrates LR digit-serial computation—operating in most-significant-digit-first (MSDF) mode—into CNN accelerator design, enabling fine-grained digit-level pipelining and parallel multiply-accumulate (MAC) operations under low interconnect overhead and small area constraints. Implemented in Verilog and synthesized in GSCL 45nm CMOS technology, DSLRCNN incorporates custom LR multipliers/adders and a digit-level pipelined convolution engine. Evaluated on AlexNet, VGG-16, and ResNet-18, it achieves 4.37×–569.11× higher peak throughput and 3.58×–44.75× improved energy efficiency (TOPS/W) over baseline accelerators, while significantly reducing inference latency, silicon area, and power consumption.

1 citationsRead paper

Genetic Fuzzy System-based Control for Final Approach of Spacecraft Rendezvous and Proximity Operations

Aug 12, 2026

This study addresses the control challenges encountered during the final approach phase of spacecraft rendezvous and docking for on-orbit servicing. To this end, an intelligent controller integrating genetic algorithms with a fuzzy inference system is proposed. The approach optimizes fuzzy rules through offline training, achieving strong generalization across diverse initial relative positions while maintaining low energy consumption. Simulation results based on an orbital dynamics model demonstrate that the designed controller reliably accomplishes the proximity operations with high efficiency and stability under varying initial conditions. Moreover, it exhibits remarkable robustness and energy-saving performance even in perturbed environments not included in the training scenarios.

0 citationsRead paper

MINT: Dynamic-Precision CNN Inference with MSDF Digit-Serial Arithmetic on FPGA

Jun 30, 2026

This work addresses the computational redundancy and energy-efficiency limitations of conventional fixed-precision CNNs on FPGAs by proposing a dynamic-precision inference accelerator based on Most-Significant-Digit-First (MSDF) serial arithmetic. The design integrates redundant signed-digit representation with a budget-constrained greedy search algorithm to dynamically select the minimal feasible integer precision (ranging from INT2 to INT7) per layer and terminate computation early once the target accuracy is achieved. To the best of our knowledge, this is the first integration of MSDF arithmetic with dynamic-precision CNN inference, enabling on-demand precision control. Implemented on a Zynq-7020 FPGA, the approach achieves throughputs of 19.86 and 18.86 GOPS and energy efficiencies of 29.51 and 26.40 GOPS/W for VGG-16 and ResNet-18, respectively, using average precisions of 5.64 and 6.04 bits—yielding over 60% higher energy efficiency than INT8 baselines with less than 2% accuracy loss.

0 citationsRead paper

A Training-Free Large Reasoning Model-based Knowledge Tracing Framework for Unified Prediction and Prescription

Jan 05, 2026arXiv.org

This work proposes Thinking-KT, a novel framework that achieves unified knowledge tracing without any model training, addressing the instability and high system complexity associated with existing approaches that rely on fine-tuning large language models (LLMs). By leveraging test-time scaling (TTS), zero-shot reasoning, and a unified output structure, Thinking-KT enables a small LLM to simultaneously perform knowledge state prediction, generate personalized feedback, and provide learning recommendations—all without fine-tuning. The method maintains competitive prediction accuracy while significantly reducing system complexity and computational overhead, achieving performance on par with fine-tuned models. These results underscore the critical role of test-time reasoning in advancing intelligent tutoring systems.

0 citationsRead paper

CURA: Size Isnt All You Need - A Compact Universal Architecture for On-Device Intelligence

Sep 29, 2025

Current on-device AI architectures face two key bottlenecks: (1) model parameters scale linearly with task complexity, and (2) poor cross-domain generalization—e.g., regression models cannot handle NLP tasks. To address these, we propose CURA, a compact, universal architecture inspired by analog audio circuits, enabling unified modeling across regression, classification, NLP, and computer vision on resource-constrained devices. Its core innovations include: (i) a network structure grounded in analog signal processing principles; (ii) lightweight parameter-sharing modules; and (iii) dynamic feature modulation for enhanced expressivity and task adaptability under extreme parameter constraints. Experiments demonstrate that CURA matches the performance of domain-specific models (up to 90% F1 score) while reducing parameter counts by up to 2500× and lowering prediction error by 1.6–2.1× across multiple benchmarks.

0 citationsRead paper

Recent publications

Latest Papers

Genetic Fuzzy System-based Control for Final Approach of Spacecraft Rendezvous and Proximity Operations

Aug 12, 2026

This study addresses the control challenges encountered during the final approach phase of spacecraft rendezvous and docking for on-orbit servicing. To this end, an intelligent controller integrating genetic algorithms with a fuzzy inference system is proposed. The approach optimizes fuzzy rules through offline training, achieving strong generalization across diverse initial relative positions while maintaining low energy consumption. Simulation results based on an orbital dynamics model demonstrate that the designed controller reliably accomplishes the proximity operations with high efficiency and stability under varying initial conditions. Moreover, it exhibits remarkable robustness and energy-saving performance even in perturbed environments not included in the training scenarios.

0 citationsRead paper

MINT: Dynamic-Precision CNN Inference with MSDF Digit-Serial Arithmetic on FPGA

Jun 30, 2026

This work addresses the computational redundancy and energy-efficiency limitations of conventional fixed-precision CNNs on FPGAs by proposing a dynamic-precision inference accelerator based on Most-Significant-Digit-First (MSDF) serial arithmetic. The design integrates redundant signed-digit representation with a budget-constrained greedy search algorithm to dynamically select the minimal feasible integer precision (ranging from INT2 to INT7) per layer and terminate computation early once the target accuracy is achieved. To the best of our knowledge, this is the first integration of MSDF arithmetic with dynamic-precision CNN inference, enabling on-demand precision control. Implemented on a Zynq-7020 FPGA, the approach achieves throughputs of 19.86 and 18.86 GOPS and energy efficiencies of 29.51 and 26.40 GOPS/W for VGG-16 and ResNet-18, respectively, using average precisions of 5.64 and 6.04 bits—yielding over 60% higher energy efficiency than INT8 baselines with less than 2% accuracy loss.

0 citationsRead paper

A Training-Free Large Reasoning Model-based Knowledge Tracing Framework for Unified Prediction and Prescription

Jan 05, 2026arXiv.org

This work proposes Thinking-KT, a novel framework that achieves unified knowledge tracing without any model training, addressing the instability and high system complexity associated with existing approaches that rely on fine-tuning large language models (LLMs). By leveraging test-time scaling (TTS), zero-shot reasoning, and a unified output structure, Thinking-KT enables a small LLM to simultaneously perform knowledge state prediction, generate personalized feedback, and provide learning recommendations—all without fine-tuning. The method maintains competitive prediction accuracy while significantly reducing system complexity and computational overhead, achieving performance on par with fine-tuned models. These results underscore the critical role of test-time reasoning in advancing intelligent tutoring systems.

0 citationsRead paper

CURA: Size Isnt All You Need - A Compact Universal Architecture for On-Device Intelligence

Sep 29, 2025

Current on-device AI architectures face two key bottlenecks: (1) model parameters scale linearly with task complexity, and (2) poor cross-domain generalization—e.g., regression models cannot handle NLP tasks. To address these, we propose CURA, a compact, universal architecture inspired by analog audio circuits, enabling unified modeling across regression, classification, NLP, and computer vision on resource-constrained devices. Its core innovations include: (i) a network structure grounded in analog signal processing principles; (ii) lightweight parameter-sharing modules; and (iii) dynamic feature modulation for enhanced expressivity and task adaptability under extreme parameter constraints. Experiments demonstrate that CURA matches the performance of domain-specific models (up to 90% F1 score) while reducing parameter counts by up to 2500× and lowering prediction error by 1.6–2.1× across multiple benchmarks.

0 citationsRead paper

Survey of AI-Powered Approaches for Osteoporosis Diagnosis in Medical Imaging

Sep 29, 2025

Osteoporosis progresses insidiously, and early radiographic identification is critical for preventing fragility fractures; however, existing AI research remains fragmented and lacks systematic integration. This paper introduces the first triaxial unified framework spanning imaging modalities (DXA, X-ray, CT, MRI), clinical tasks (risk prediction, diagnosis, subtyping), and AI methodologies (classical ML, CNNs, Transformers, self-supervised learning, XAI), underpinned by a PRISMA-guided systematic review of 127 studies. Key contributions include: (1) construction of a domain-specific knowledge graph and a technology roadmap; (2) identification of three critical bottlenecks—data scarcity, insufficient external validation, and limited interpretability; and (3) establishment of an interdisciplinary collaboration paradigm for AI developers, radiologists, and clinicians to advance precision early screening and individualized management of osteoporosis.

0 citationsRead paper