arXiv subjects

Mikolaj Walczak

Publications and source records attributed to Mikolaj Walczak.

5 recordsLinked to original sources

arXiv2026

ShatterQuant: Breaking Uniform Precision with Block-Wise Mixed-Precision on a Systolic Transformer Hardware Accelerator

Due to limited support for intra-tensor heterogeneous precision in conventional accelerators, neural network quantization remains largely restricted to per-tensor precision assignment. We present ShatterQuant, a hardware-software co-designed framework enabling mixed-precision quantization within each tensor by assigning independent bit-widths to blocks of a weight projection. ShatterQuant couples precision granularity with PE configuration, such that each precision determines an effective block height. We introduce (1) a hardware-aware post-training method that assigns intra-tensor precision based on block-level standard deviation and weight sensitivity; (2) the ShatterQuant Transformer Accelerator supporting 1/2/4/8-bit weight precision, precision-dependent PE configuration, block rescaling, and integrated softmax and piecewise-linear nonlinearities; and (3) an evaluation of model-hardware tradeoffs using an implementation in the TSMC 16nm PDK operating at 1 GHz, achieving 1.5 TOPS, 760 GOPS/$mm^2$ area efficiency, and 2.8 TOPS/W energy efficiency. On DeiT and ImageNet-1K, ShatterQuant achieves accuracy within $3.3\%$ of state-of-the-art mixed-precision techniques while using a 2 bit lower effective bitwidth, while for PixelDiT demonstrates comparable generation quality. ShatterQuant demonstrates how fine-grained intra-tensor mixed-precision can be realized through hardware-software co-design.

cs.AR↗

arXiv2025

Invited Paper: BitMedViT: Ternary-Quantized Vision Transformer for Medical AI Assistants on the Edge

Vision Transformers (ViTs) have demonstrated strong capabilities in interpreting complex medical imaging data. However, their significant computational and memory demands pose challenges for deployment in real-time, resource-constrained mobile and wearable devices used in clinical environments. We introduce, BiTMedViT, a new class of Edge ViTs serving as medical AI assistants that perform structured analysis of medical images directly on the edge. BiTMedViT utilizes ternary- quantized linear layers tailored for medical imaging and com- bines a training procedure with multi-query attention, preserving stability under ternary weights with low-precision activations. Furthermore, BiTMedViT employs task-aware distillation from a high-capacity teacher to recover accuracy lost due to extreme quantization. Lastly, we also present a pipeline that maps the ternarized ViTs to a custom CUDA kernel for efficient memory bandwidth utilization and latency reduction on the Jetson Orin Nano. Finally, BiTMedViT achieves 86% diagnostic accuracy (89% SOTA) on MedMNIST across 12 datasets, while reducing model size by 43x, memory traffic by 39x, and enabling 16.8 ms inference at an energy efficiency up to 41x that of SOTA models at 183.62 GOPs/J on the Orin Nano. Our results demonstrate a practical and scientifically grounded route for extreme-precision medical imaging ViTs deployable on the edge, narrowing the gap between algorithmic advances and deployable clinical tools.

eess.IV↗

arXiv2025

EDEN: Entorhinal Driven Egocentric Navigation Toward Robotic Deployment

Deep reinforcement learning agents are often fragile while humans remain adaptive and flexible to varying scenarios. To bridge this gap, we present EDEN, a biologically inspired navigation framework that integrates learned entorhinal-like grid cell representations and reinforcement learning to enable autonomous navigation. Inspired by the mammalian entorhinal-hippocampal system, EDEN allows agents to perform path integration and vector-based navigation using visual and motion sensor data. At the core of EDEN is a grid cell encoder that transforms egocentric motion into periodic spatial codes, producing low-dimensional, interpretable embeddings of position. To generate these activations from raw sensory input, we combine fiducial marker detections in the lightweight MiniWorld simulator and DINO-based visual features in the high-fidelity Gazebo simulator. These spatial representations serve as input to a policy trained with Proximal Policy Optimization (PPO), enabling dynamic, goal-directed navigation. We evaluate EDEN in both MiniWorld, for rapid prototyping, and Gazebo, which offers realistic physics and perception noise. Compared to baseline agents using raw state inputs (e.g., position, velocity) or standard convolutional image encoders, EDEN achieves a 99% success rate, within the simple scenarios, and >94% within complex floorplans with occluded paths with more efficient and reliable step-wise navigation. In addition, as a replacement of ground truth activations, we present a trainable Grid Cell encoder enabling the development of periodic grid-like patterns from vision and motion sensor data, emulating the development of such patterns within biological mammals. This work represents a step toward biologically grounded spatial intelligence in robotics, bridging neural navigation principles with reinforcement learning for scalable deployment.

cs.RO↗

arXiv2025

ATLASv2: LLM-Guided Adaptive Landmark Acquisition and Navigation on the Edge

Autonomous systems deployed on edge devices face significant challenges, including resource constraints, real-time processing demands, and adapting to dynamic environments. This work introduces ATLASv2, a novel system that integrates a fine-tuned TinyLLM, real-time object detection, and efficient path planning to enable hierarchical, multi-task navigation and manipulation all on the edge device, Jetson Nano. ATLASv2 dynamically expands its navigable landmarks by detecting and localizing objects in the environment which are saved to its internal knowledge base to be used for future task execution. We evaluate ATLASv2 in real-world environments, including a handcrafted home and office setting constructed with diverse objects and landmarks. Results show that ATLASv2 effectively interprets natural language instructions, decomposes them into low-level actions, and executes tasks with high success rates. By leveraging generative AI in a fully on-board framework, ATLASv2 achieves optimized resource utilization with minimal prompting latency and power consumption, bridging the gap between simulated environments and real-world applications.

cs.RO↗

arXiv2022

A Reliable and Low Latency Synchronizing Middleware for Co-simulation of a Heterogeneous Multi-Robot Systems

Search and rescue, wildfire monitoring, and flood/hurricane impact assessment are mission-critical services for recent IoT networks. Communication synchronization, dependability, and minimal communication jitter are major simulation and system issues for the time-based physics-based ROS simulator, event-based network-based wireless simulator, and complex dynamics of mobile and heterogeneous IoT devices deployed in actual environments. Simulating a heterogeneous multi-robot system before deployment is difficult due to synchronizing physics (robotics) and network simulators. Due to its master-based architecture, most TCP/IP-based synchronization middlewares use ROS1. A real-time ROS2 architecture with masterless packet discovery synchronizes robotics and wireless network simulations. A velocity-aware Transmission Control Protocol (TCP) technique for ground and aerial robots using Data Distribution Service (DDS) publish-subscribe transport minimizes packet loss, synchronization, transmission, and communication jitters. Gazebo and NS-3 simulate and test. Simulator-agnostic middleware. LOS/NLOS and TCP/UDP protocols tested our ROS2-based synchronization middleware for packet loss probability and average latency. A thorough ablation research replaced NS-3 with EMANE, a real-time wireless network simulator, and masterless ROS2 with master-based ROS1. Finally, we tested network synchronization and jitter using one aerial drone (Duckiedrone) and two ground vehicles (TurtleBot3 Burger) on different terrains in masterless (ROS2) and master-enabled (ROS1) clusters. Our middleware shows that a large-scale IoT infrastructure with a diverse set of stationary and robotic devices can achieve low-latency communications (12% and 11% reduction in simulation and real) while meeting mission-critical application reliability (10% and 15% packet loss reduction) and high-fidelity requirements.

cs.RO↗