This study surveys emerging AI agent architectures for reasoning, planning, and tool calling, revealing their capabilities and limitations.

Tula Masterman, Sandi Besen, Mason Sawtell et al.

Language Ranker quantifies multilingual performance by comparing internal Transformer representations with an English baseline, revealing resource-based disparities.

Zihao Li, Yucheng Shi, Zirui Liu et al.

This study analyzes automatic metrics for meeting summarization, revealing their limited sensitivity to meeting-specific errors and model architecture differences.

Frederic Kirstein, Jan Philip Wahle, Terry Ruas et al.

Proposes LAPTOP-Diff, combining layer pruning and normalized distillation to compress diffusion models with only 4% performance loss at 50% pruning.

Dingkun Zhang, Sijia Li, Chen Chen et al.

Semantics-Aware Attention Guidance (SAG) improves cancer diagnosis accuracy on whole slide images by integrating tissue and cellular cues, boosting performance by over 6%.

Kechun Liu, Wenjun Wu, Joann G. Elmore et al.

Proposes FAC framework with context-specific independence and JACI algorithm for automatic causal discovery in complex environments.

Caleb Chuck, Sankaran Vaidyanathan, Stephen Giguere et al.

MiniCheck achieves GPT-4 level fact-checking using synthetic data, reducing costs by 400x.

Liyan Tang, Philippe Laban, Greg Durrett

SPAG enhances LLM reasoning by 5% through self-playing adversarial language games.

Pengyu Cheng, Tianhao Hu, Han Xu et al.

This study evaluates VLMs' ability to infer private attributes from inconspicuous images, achieving up to 77.6% accuracy, highlighting privacy risks.

Batuhan Tömekçe, Mark Vero, Robin Staab et al.

MK-SGN combines multimodal fusion and knowledge distillation for skeleton-based action recognition, achieving 98% energy reduction.

Naichuan Zheng, Hailun Xia, Zeyu Liang et al.

Proposed a multi-message shuffle protocol achieving optimal error with O(min(nε², d)) messages; proved lower bounds and analyzed robustness issues.

Hilal Asi, Vitaly Feldman, Jelani Nelson et al.

Introduces MMInA benchmark for evaluating multihop multimodal web agents, using 1050 real evolving websites to assess long-range reasoning.

Shulin Tian, Ziniu Zhang, Liangyu Chen et al.

Ctrl-Adapter efficiently adapts ControlNet, achieving superior performance on COCO and DAVIS 2017.

Han Lin, Jaemin Cho, Abhay Zala et al.

Large language models like GPT-4 exhibit self-recognition, with a linear link to their self-preference bias, validated through fine-tuning experiments.

Arjun Panickssery, Samuel R. Bowman, Shi Feng

Join-aware multi-table retrieval using mixed-integer programming improves F1 by 9.3%.

Peter Baile Chen, Yi Zhang, Dan Roth

Achieved 90% of hand-written code performance using MLIR-based AI compiler.

Renato Golin, Lorenzo Chelini, Adam Siemieniuk et al.

SparseOcc employs sparse latent representations for semantic occupancy prediction, reducing FLOPs by 74.9% and increasing mIoU from 12.8% to 14.1%.

Pin Tang, Zhongdao Wang, Guoqing Wang et al.

Wasserstein Wormhole, a transformer-based autoencoder, embeds distributions into a latent space where Euclidean distances approximate Wasserstein distances, enabling linear-time OT computations for large datasets.

Doron Haviv, Russell Zhang Kunes, Thomas Dougherty et al.

The paper introduces ATACOM, a method for safe reinforcement learning on constraint manifolds, effectively applied to a robot air hockey task.

Puze Liu, Haitham Bou-Ammar, Jan Peters et al.

This study probes large-scale visual models' 3D awareness via task-specific probes, revealing significant limitations in depth, normals, and multiview consistency.

Mohamed El Banani, Amit Raj, Kevis-Kokitsi Maninis et al.