Papers by Vineet Gandhi
Comprehensive Multi-Modal Interactions for Referring Image Segmentation (2022.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for RIS compute different forms of interactions sequentially or ignore intra-modal interactions. |
| Approach: | They propose a method which outputs a segmentation map corresponding to the natural language description. |
| Outcome: | The proposed method performs on four benchmark datasets and shows significant performance gains over the existing state-of-the-art methods. |
Major Entity Identification: A Generalizable Alternative to Coreference Resolution (2024.emnlp-main)
Copied to clipboard
| Challenge: | Prior work identified annotation differences as one of the main reasons for the limited generalization gap in coreference resolution models. |
| Approach: | They propose an alternative referential task where the target entities are assumed to be specified in the input and the task is limited to the frequent entities. |
| Outcome: | The proposed model generalizes well across domains on multiple datasets with supervised models and LLM-based few-shot prompting. |
Empathic Machines: Using Intermediate Features as Levers to Emulate Emotions in Text-To-Speech Systems (2022.naacl-main)
Copied to clipboard
| Challenge: | a method to control affective prosody of text-to-speech systems is proposed to use phoneme-level intermediate features as levers . DS is used to disentangle features relating to affective proody from those due to acoustics conditions and speaker identity . |
| Approach: | They propose a method to control the emotional prosody of Text to Speech systems by using phoneme-level intermediate features as levers. |
| Outcome: | The proposed method improves over the prior art in emulating emotion in speech . it adds the much-coveted "human touch" in machine dialogue, the authors say . |
ParrotTTS: Text-to-speech synthesis exploiting disentangled self-supervised representations (2024.findings-eacl)
Copied to clipboard
| Challenge: | ParrotTTS can train a multi-speaker variant using transcripts from a single speaker in low resource setup and generalizes to languages not seen while training the self-supervised backbone. |
| Approach: | They propose a modular text-to-speech synthesis model that can train a multi-speaker variant using transcripts from a single speaker. |
| Outcome: | The proposed model outperforms state-of-the-art multi-lingual text-to-speech models using only a fraction of paired data as latter. |
IdentifyMe: A Challenging Long-Context Mention Resolution Benchmark for LLMs (2025.naacl-short)
Copied to clipboard
| Challenge: | Recent evaluations of LLMs on coreference resolution have revealed that traditional output formats and evaluation metrics do not fully capture the models’ referential understanding. |
| Approach: | They propose a benchmark for mention resolution presented in a multiple-choice question format and a curated mixture of different mention types and corresponding entities. |
| Outcome: | The proposed model achieves 81.9% accuracy while the open model achieve 80%. |