Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile
Articles by Jay Mohta
Do VLMs read or rewrite?
Vision Language Models (VLMs) are increasingly adopted for document understanding tasks, often replacing traditional OCR systems. However, VLMs exhibit a fundamental difference: they frequently correct or rewrite imperfect text rather than transcribe it literally, a behavior that remains largely underexplored. We present a systematic investigation through controlled experiments with intentionally perturbed text across seven models and 1,706 documents.
Amazon Science at CVPR 2026
CVPR 2026 Workshop on Grounded Retrieval and Agentic Intelligence for Vision-Language (GRAIL-V) June 3 Website: Link Accepted Papers: "CoCoA-DVC: Consistency and Concept Aware Training for Dense Video Captioning", presented by Jay Nitin Paranjape, Yue Guo, Sankar Venkataraman, Vishal M.
ReScorer: An aggregation and alignment technique for building trust into LLM reasons
Large language models (LLMs) offer substantial potential for automating labeling tasks, showcasing robust zero-shot performance across diverse classification tasks. The LLM-generated reasons that accompany these classifications contain signals about the quality of the classifications. Estimates of quality of these reasons can, in essence, be used to detect potentially incorrect predictions.
Aligning vision language models with contrastive learning
In recent years, Vision Language Models (VLMs) have achieved significant advancements due to the success of large language models. The common strategy for aligning vision and language models involves a two-step process: an alignment (or pretraining) stage and an instruction tuning stage. During the alignment stage, a projection module is trained to map image embeddings into the language space using a paired image-text dataset.
Are large language models good annotators?
Numerous Natural Language Processing (NLP) tasks require precisely labeled data to ensure effective model training and achieve optimal performance. However, data annotation is marked by substantial costs and time requirements, especially when requiring specialized domain expertise or annotating a large number of samples. In this study, we investigate the feasibility of employing large language models (LLMs) as replacements for human annotators.
Prompting language models improves performance in imbalanced setting
“Find a coffee shop between here and San Francisco,” “How much longer is Target on main street open?,” “What is the next train to work,” “I need directions to the nearest gas station;” these are some of the utterances answered by the Alexa Local Information team. We enable customers to search for places, products, or services around them, and provide personalized travel, commute and navigation experiences to the places they find.
Show More
loading
Actions
Is this you?
As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.Get in touch with Jay
Contact Jay, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack