arXiv subjects

Tapas Kumar Maiti

Publications and source records attributed to Tapas Kumar Maiti.

5 recordsLinked to original sources

arXiv2026

ChainLearn: A Blockchain-Based Capacity-Aware Framework for Federated Ensemble Learning

Federated learning is used in medical imaging where privacy prohibits centralizing data. Standard federated algorithms assume homogeneous hardware, identical architectures, and centralized aggregation, which fails when hospitals have unequal compute resources. We propose capacity-aware coordination: measure each hospital's throughput, assign capacity-appropriate architectures (MobileNetV3-Small, EfficientNet-B0, ResNet-50), and combine predictions via weighted ensemble. Weak and strong hospitals can participate without forcing uniform architectures. We separate on-chain policy from off-chain learning. A Solidity contract stores hospital registration, benchmark hashes, metrics, and weights. Hospitals train locally and submit only hashes and scalars (not parameters). Weighted ensemble inference is computed off-chain. Experiments on PneumoniaMNIST and DermaMNIST (5 seeds, 3 non-IID levels) show our method achieves lower or equal calibration error versus equal-weight ensemble and competitive accuracy versus FedAvg, FedProx, and FedMD. Communication overhead is 224 bytes per round, a reduction of over 912,000x compared to FedAvg.

cs.LG↗

arXiv2025

SpectralKrum: A Spectral-Geometric Defense Against Byzantine Attacks in Federated Learning

Federated Learning (FL) distributes model training across clients who retain their data locally, but this architecture exposes a fundamental vulnerability: Byzantine clients can inject arbitrarily corrupted updates that degrade or subvert the global model. While robust aggregation methods (including Krum, Bulyan, and coordinate-wise defenses) offer theoretical guarantees under idealized assumptions, their effectiveness erodes substantially when client data distributions are heterogeneous (non-IID) and adversaries can observe or approximate the defense mechanism. This paper introduces SpectralKrum, a defense that fuses spectral subspace estimation with geometric neighbor-based selection. The core insight is that benign optimization trajectories, despite per-client heterogeneity, concentrate near a low-dimensional manifold that can be estimated from historical aggregates. SpectralKrum projects incoming updates into this learned subspace, applies Krum selection in compressed coordinates, and filters candidates whose orthogonal residual energy exceeds a data-driven threshold. The method requires no auxiliary data, operates entirely on model updates, and preserves FL privacy properties. We evaluate SpectralKrum against eight robust baselines across seven attack scenarios on CIFAR-10 with Dirichlet-distributed non-IID partitions (alpha = 0.1). Experiments spanning over 56,000 training rounds show that SpectralKrum is competitive against directional and subspace-aware attacks (adaptive-steer, buffer-drift), but offers limited advantage under label-flip and min-max attacks where malicious updates remain spectrally indistinguishable from benign ones.

cs.LG↗

arXiv2024

VLSI Architectures of Forward Kinematic Processor for Robotics Applications

This paper aims to get a comprehensive review of current-day robotic computation technologies at VLSI architecture level. We studied several repots in the domain of robotic processor architecture. In this work, we focused on the forward kinematics architectures which consider CORDIC algorithms, VLSI circuits of WE DSP16 chip, parallel processing and pipelined architecture, and lookup table formula and FPGA processor. This study gives us an understanding of different implementation methods for forward kinematics. Our goal is to develop a forward kinematics processor with FPGA for real-time applications, requires a fast response time and low latency of these devices, useful for industrial automation where the processing speed plays a great role.

cs.RO↗

arXiv2019

On-Chip Concentration and Patterning of Biological Cells Using Interplay of Electrical and Thermal Fields

We demonstrate a method of concentrating and patterning of biological cells on a chip, exploiting the confluence of electric and thermal fields, without necessitating the use of any external heating or illuminating source. The technique simply employs two parallel plate electrodes and an insulating layer over the bottom electrode, with a drilled insulating layer for inducing localized variations in the thermal field. A strong induced electric field, in the process, penetrates through the narrow hole and generates highly non-uniform heating, which in turn, results in gradients in electrical properties and induces mobile charges to impose directional fluid flow. The toroidal vortices, induced by secondary electrokinetic forces originating out of temperature-dependent electrical property variations, transport the suspended cells towards a hot-spot site of the chip, for rapid concentrating and patterning into different shaped clusters based on pre-designed conditions, without exceeding safe temperature limits that do not result in damage of thermally labile biological samples. We characterize the efficacy of the cell trapping process for two different biological entities, namely, Escherichia coli bacteria and yeast cell. These results may be of profound importance towards developing novel biomedical microdevices for drug discovery, antibiotic resistance assessment and medical diagnostics.

physics.bio-ph↗

arXiv2016

A study in blood at pinch

The complex fluidic nature of blood, though necessary to serve different physiological purposes, gives rise to daunting challenges in developing unified conceptual paradigm describing the underlying physics of blood at pinch, which may otherwise be essential for understanding various bio-technological processes demanding precise and efficient handling of blood samples. Intuitively, a blood-drop may be formed simply by dripping. However, the pinch-off dynamics leading to blood-drop-breakup is elusively more complex than what may be portrayed by any unique model depicting the underlying morpho-dynamics, as our study reveals. With blood samples, here we observe two distinctive modes of the breakup process. One mode corresponds to incessant collapsing of a liquid-neck, while in other mode formation and thinning of an extended long thread leads to the breakup and drop formation. We further show that these modes are respectively described by power law and exponential law based universal scaling dynamics, depicting the temporal evolution of the neck thickness of the blood filament. Our results are likely to bear far-reaching consequences in diverse applications, ranging from forensic sciences to droplet based microfluidic technology.

physics.flu-dyn↗