Posters

New method for eigenvector estimation achieves optimal rates without coherence dependence, improving low-rank matrix denoising and related tasks.

CoFie: A novel local geometry-aware neural surface representation dramatically improves accuracy and efficiency in 3D shape modeling by using coordinate fields to compress local shape information.

CORY: a novel multi-agent RL framework boosts LLM fine-tuning!

Code Rosetta pushes the boundaries of unsupervised code translation by introducing the first encoder-decoder model that efficiently translates between programming languages and their parallel HPC exte…

LeTCC: A novel learning-theoretic framework for resilient distributed computing, achieving faster convergence and higher accuracy than existing methods by integrating learning theory principles with c…

CODE combats LMM hallucinations by contrasting self-generated descriptions with visual content during decoding, enhancing response accuracy without retraining.

New program synthesis method, REX, leverages Thompson Sampling to balance exploration and exploitation in iterative LLM code refinement, solving more problems with fewer model calls.

CODA: A novel modeling scheme tackles subpopulation shifts in machine learning by disentangling spurious correlations, augmenting data strategically, and using reweighted consistency loss for improved…

CoBo: A novel bilevel optimization algorithm for collaborative learning surpasses existing methods by efficiently selecting helpful clients, resulting in superior performance and scalability.

Hierarchical concept bottleneck models boost interpretability and accuracy in visual classification by uncovering both high-level and low-level concepts.

Researchers developed CNCA, a novel method that generates realistic and customizable adversarial camouflage for vehicle detectors by leveraging a pre-trained diffusion model, surpassing existing metho…

This paper introduces novel non-adaptive algorithms for clustering using subset queries, achieving near-linear query complexity and improving upon existing limitations of pairwise query methods.

RecPiece: Relational Clustering for Better Knowledge Graph Embedding Anchors

Researchers strengthen understanding of transformer self-attention by proving asymptotic convergence to single clusters under causal masking, linking it to the Rényi parking problem.

Cluster-Learngene efficiently initializes elastic-scale Vision Transformers by adaptively clustering and inheriting key modules from a large ancestry model, saving resources and boosting downstream ta…

CLUES: Collaborative learning selects high-quality private data for LLM fine-tuning via training dynamics, significantly boosting performance in diverse domains.

COIN: A novel method for Cloud Object Detector Adaptation that integrates knowledge from cloud models and CLIP to train highly accurate target detectors, achieving state-of-the-art performance.

CLOVER: A closed-loop visuomotor framework using generative visual plans & feedback mechanisms achieves state-of-the-art results in long-horizon robotic manipulation tasks.

CLIPCEIL enhances CLIP’s domain generalization by refining feature channels for domain invariance and aligning image-text embeddings, achieving state-of-the-art performance.

MirrorCLIP disentangles text from images in CLIP using mirror reflection differences, enhancing robustness against text-visual image confusion.