Searcharxiv⌕ Search

arXiv subjects

Kalyan Bhetwal

Publications and source records attributed to Kalyan Bhetwal.

3 recordsLinked to original sources

UnSPAD: Uncertainty-aware Single Photon Burst Photography

We introduce UnSPAD, an uncertainty-aware burst photography framework capable of fusing and denoising arbitrarily long sequences of binary single photon avalanche diode (SPAD) measurements. Unlike prior methods that require fixed burst lengths or rely on powerful generative priors, UnSPAD combines learned motion estimation with uncertainty-informed measurement fusion and denoising to produce temporally consistent, hallucination-free reconstructions. UnSPAD first forms an initial scene estimate by averaging flow-aligned binary frames. UnSPAD then uses per-pixel uncertainty maps to perform uncertainty-aware denoising; the denoiser is informed about which regions accumulated the fewest observations (and thus contain the most relative shot noise) due to motion, occlusion, or alignment errors. This two-step reconstruction pipeline helps mitigate hallucinations---outside of the structure implicitly introduced by the flow estimates, the initial scenes estimates are entirely prior-free. Experiments on real and simulated SPAD and quantum image sensor (QIS) benchmarks demonstrate state-of-the-art reconstruction quality, improved temporal consistency, and superior color fidelity relative to the state of the art.

cs.CV↗

Name-Bias in Large Language Models

Names carry demographic indicators such as age, gender, and race, which makes them a subtle but strong target of bias in large language models (LLMs). In this work, we present an evaluation of name-based bias across 6 local large language models spanning multiple providers and parameter scales. Using a fill-in-the-blank task, we prompt these models to choose between pairs of first names drawn from U.S. Social Security Administration datasets and a race classification dataset by Rosenman et al.. Our experimental de- sign isolates age and gender effects through balanced name pairings and sentence templates that share a fixed syntactic structure. Across 300,000 local LLM responses, we find consistent and measurable name bias. Models prefer names that are less stereotypically male, older, popular, and white. Furthermore, we find that name-bias is highly sentence-dependent, but not more than associated stereotypes, and gender is the most significant bias, followed by age, name popularity, and race. Our results demonstrate that name-based bias remains pervasive even in local LLMs, highlighting the necessity for standardized evaluation methods and mitigation strategies.

cs.CL↗

Analysis of Software Engineering Practices in General Software and Machine Learning Startups

Context: On top of the inherent challenges startup software companies face applying proper software engineering practices, the non-deterministic nature of machine learning techniques makes it even more difficult for machine learning (ML) startups. Objective: Therefore, the objective of our study is to understand the whole picture of software engineering practices followed by ML startups and identify additional needs. Method: To achieve our goal, we conducted a systematic literature review study on 37 papers published in the last 21 years. We selected papers on both general software startups and ML startups. We collected data to understand software engineering (SE) practices in five phases of the software development life-cycle: requirement engineering, design, development, quality assurance, and deployment. Results: We find some interesting differences in software engineering practices in ML startups and general software startups. The data management and model learning phases are the most prominent among them. Conclusion: While ML startups face many similar challenges to general software startups, the additional difficulties of using stochastic ML models require different strategies in using software engineering practices to produce high-quality products.

cs.SE↗