A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Adam Fisch, Shubhendu Trivedi, Fantine Huot, et al.
This paper formalizes AI model routing with costly value estimation as Pandora's Box, introducing closed-form policies that decide when refining value estimates is worth the cost.
Xu Cao, Houze Yang, Vipin Gunda, et al.
Introduces Promptable Gaze Target Estimation (PGE), a concept-driven paradigm using text or visual prompts for end-to-end gaze analysis, with a new dataset Gaze-Co and model GazeAnywhere.
Unknown
This paper proposes a minimalist offline RL approach that avoids out-of-distribution value overestimation by constraining the policy to the data support, achieving strong performance with simple modifications.
L. Yang, Zhilong Zhang, Yang Song, et al.
A comprehensive survey of diffusion models, categorizing research into efficient sampling, improved likelihood estimation, and handling special data structures, with a review of applications across domains.
Jinsen Su, Yongdong Luo, Yuexiao Ma, et al.
OmniScope is a training-free token compression framework for omnimodal LLMs that decouples audio and video relevance estimation using the query as a shared anchor, achieving up to 3.53x prefill speedup with minimal accuracy loss.
Rahul Raguram, Ondřej Chum, Marc Pollefeys, et al.
USAC extends RANSAC with a universal framework integrating advanced sampling, verification, and model refinement for robust estimation.
Jakob Gawlikowski, Cedrique Rovile Njieutcheu Tassi, Mohsin Ali, et al.
A comprehensive survey of uncertainty estimation in deep neural networks, covering sources, modeling approaches, calibration, and practical applications.
Michael I. Jordan, Robert A. Jacobs
Introduces a tree-structured supervised learning architecture using hierarchical mixtures of experts with an EM algorithm for parameter estimation.
Ioannis Patras, Ella Hendriks, Reginald L. Lagendijk
This paper proposes a spatio-temporal video segmentation method that labels watershed segments using MAP estimation with a Markov random field model.
Jerome H. Friedman
This paper introduces gradient boosting machines, a general paradigm for function estimation via stagewise additive expansions and steepest-descent minimization in function space.
Hado van Hasselt, Arthur Guez, David Silver
This paper shows that DQN overestimates action values in Atari games and proposes a Double DQN algorithm that reduces overestimation and improves performance.
Unknown
GenPRM scales test-time compute of process reward models via generative reasoning and improved label estimation.