6 papers
SCRIBE: Diagnostic Evaluation and Rich Transcription Models for Indic ASR
Kavya Manohar, Arghya Bhattacharya, Kush Juvekar +1
Automatic speech recognition replaces typing only when correction costs less than manual entry - a threshold determined by error types, not counts: fixing a misrecognized domain te…
Vividh-ASR: A Complexity-Tiered Benchmark and Optimization Dynamics for Robust Indic Speech Recognition
Kush Juvekar, Kavya Manohar, Aditya Srinivas Menon +2
Fine-tuning multilingual ASR models like Whisper for low-resource languages often improves read speech but degrades spontaneous audio performance. To diagnose this mismatch, we int…
CourtNav: Voice-Guided, Anchor-Accurate Navigation of Long Legal Documents in Courtrooms
Sai Khadloya, Kush Juvekar, Arghya Bhattacharya +1
Judicial work depends on close reading of long records, charge sheets, pleadings, annexures, orders, often spanning hundreds of pages. With limited staff support, exhaustive readin…
Are LLMs Court-Ready? Evaluating Frontier Models on Indian Legal Reasoning
Kush Juvekar, Arghya Bhattacharya, Sai Khadloya +1
Large language models (LLMs) are entering legal workflows, yet we lack a jurisdiction-specific framework to assess their baseline competence therein. We use India's public legal ex…
Introducing a new hyper-parameter for RAG: Context Window Utilization
Kush Juvekar, Anupam Purwar
This paper introduces a new hyper-parameter for Retrieval-Augmented Generation (RAG) systems called Context Window Utilization. RAG systems enhance generative models by incorporati…
COS-Mix: Cosine Similarity and Distance Fusion for Improved Information Retrieval
Kush Juvekar, Anupam Purwar
This study proposes a novel hybrid retrieval strategy for Retrieval-Augmented Generation (RAG) that integrates cosine similarity and cosine distance measures to improve retrieval p…