Debottam Dutta
Scholar

Debottam Dutta

Google Scholar ID: gBYk86YAAAAJ
University of Illinois at Urbana-Champaign
Deep learningGenerative modelsSpeech processing
Homepage↗Google Scholar↗
Citations & Impact
All-time
Citations
142
 
H-index
5
 
i10-index
2
 
Publications
15
 
Co-authors
0
 
Contact
Emaildebottam.dutta7@gmail.comCVOpen ↗GitHubOpen ↗LinkedInOpen ↗
Publications
5 items
Personalized Image Generation via Human-in-the-loop Bayesian Optimization
2026
Cited
0
CO3: Contrasting Concepts Compose Better
2025
Cited
0
Learning Energy-based Variational Latent Prior for VAEs
2025
Cited
0
Estimating Multi-chirp Parameters using Curvature-guided Langevin Monte Carlo
2025
Cited
0
Multi-Source Music Generation with Latent Diffusion
arXiv.org · 2024
Cited
2
Resume
Academic Achievements
  • 1. Paper 'Estimating Multi-chirp Parameters using Curvature-guided Langevin Monte Carlo' accepted at ICASSP 2025.
  • 2. A paper accepted at NeurIPS 2024 Workshop on AI-Driven Speech, Music, and Sound Generation.
  • 3. Journal paper accepted for TASLP.
  • 4. Coswara dataset paper accepted at Nature Scientific Data.
  • 5. Two papers accepted at Interspeech 2022.
Research Experience
  • Prior to joining the PhD program, worked as a Research Fellow at LEAP Lab, Indian Institute of Science, under Prof. Sriram Ganapathy, focusing on learning useful and interpretable audio representations for different downstream tasks.
Education
  • PhD Student in ECE at University of Illinois at Urbana-Champaign, under the guidance of Prof. Romit Roy Choudhury.
Background
  • Research interests: Generative models and their applications to Audio and Images. Currently working on making text-to-image generation more robust.