Aneesh Komanduri Data-verified
Affiliation confirmed via AI analysis of OpenAlex, ORCID, and web sources.
Researcher
unknown
Research Areas
Biography and Research Information
OverviewAI-generated summary
Aneesh Komanduri's research focuses on advancing machine learning techniques, particularly within the domains of artificial intelligence and computer vision. His work has explored the development of novel methods for learning identifiable causal representations, as demonstrated in his publication "SCM-VAE: Learning Identifiable Causal Representations via Structural Knowledge." Komanduri has also investigated applications in sentiment recognition, proposing "Counterfactual Thinking Driven Emotion Regulation for Image Sentiment Recognition." His research further extends to graph-based learning with "Neighborhood Random Walk Graph Sampling for Regularized Bayesian Graph Convolutional Neural Networks." More recently, his work addresses challenges in large vision-language models, including "Cross-Modal Attention Guided Unlearning in Vision-Language Models" and the development of benchmarks for visual causal reasoning, such as "CausalVLBench: Benchmarking Visual Causal Reasoning in Large Vision-Language Models." Komanduri's scholarship includes 6 publications and has garnered 14 citations, with an h-index of 2.
Metrics
- h-index: 3
- Publications: 7
- Citations: 37
Selected Publications
-
CausalVLBench: Benchmarking Visual Causal Reasoning in Large Vision-Language Models (2025)
-
Toward Causal Generative Modeling: From Representation to Generation (2025)
-
Causal Diffusion Autoencoders: Toward Counterfactual Generation via Diffusion Probabilistic Models (2024)
-
Counterfactual Thinking Driven Emotion Regulation for Image Sentiment Recognition (2024)
-
SCM-VAE: Learning Identifiable Causal Representations via Structural Knowledge (2022)
-
Neighborhood Random Walk Graph Sampling for Regularized Bayesian Graph Convolutional Neural Networks (2021)
Collaborators
Researchers in the database who share publications
Similar Researchers
Based on overlapping research topics