Esaú Villatoro-Tello

Context Projector: Complementary Keyword and Dialogue Context Embeddings for LLM-based ASR

Closing the Speech-Text Gap with Limited Audio for Effective Domain Adaptation in LLM-Based ASR

MB using only 10% of the target-domain speech achieves word error rates comparable to, or better than, conventional ASR fine-tuning with the full dataset, indicating that small amounts of speech provide a strong modality-alignment signal.

Rethinking Human-Aligned Evaluation: An Analysis of Semantic Metrics Beyond WER

Word Error Rate (WER), the most commonly used metric for Automatic Speech Recognition (ASR), treats every lexical deviation from the reference as equally costly, regardless of whether it changes meaning. This raises the question: does WER actually track how humans judge ASR transcript quality? We introduce HATS-en, an...

When Consistency Becomes Bias: Interviewer Effects in Semi-Structured Clinical Interviews

This work analyzes three datasets: ANDROIDS, DAIC-WOZ, E-DAIC and identifies a systematic bias from interviewer prompts in semi-structured interviews that highlights a cross-dataset, architecture-agnostic bias and emphasizes the need for analyses that localize decision evidence by time and speaker to ensure models lear...

Reading Emotions in the Token Space: Discriminative Adaptation of SpeechLLMs for Emotion Recognition

A discriminative adaptation that reads the final prompt token's hidden state through a classification head, producing a label in one forward pass without modifying the backbone, gives a controlled comparison of generative and discriminative inference in an otherwise identical speechLLM.

When Synthetic Speech Is All You Have: Better Call GRPO

This work shows that Group Relative Policy Optimization (GRPO) extracts far more from the same synthetic speech than SFT, and traces the gain to behavior rather than representation: GRPO reduces insertion errors by improving stopping calibration and speech-to-text alignment by better anchoring attention to audio, leavi...