1. 9.3 Sequential Decision Making with Expert Demonstrations under Unobserved Heterogeneity
   • Authors: Vahid Balazadeh, Keertana Chidambaram, Viet Nguyen, Rahul G. Krishnan, Vasilis Syrgkanis
   • Reason: The problem addressed is highly relevant in practical domains like healthcare and finance, and the approach is innovative, using a zero-shot meta-reinforcement learning framework. The involvement of authors with expertise in machine learning and economics may indicate substantial impact.
2. 9.2 Structured Reinforcement Learning for Media Streaming at the Wireless Edge
   • Authors: Archana Bura, Sarat Chandra Bobbili, Shreyas Rameshkumar, Desik Rengarajan, Dileep Kalathil, Srinivas Shakkottai
   • Reason: The paper addresses the important application of media streaming over wireless networks using CMDP and presents a real-world evaluation. The structured learning approach with empirical validation suggests a potentially substantial influence in practical settings.
3. 8.9 AdaDemo: Data-Efficient Demonstration Expansion for Generalist Robotic Agent
   • Authors: Tongzhou Mu, Yijie Guo, Jie Xu, Ankit Goyal, Hao Su, Dieter Fox, Animesh Garg
   • Reason: This work is at the intersection of imitation learning and robotics, and it tackles the challenge of scaling up demonstrations. The presence of prominent authors in robotics and the evaluation across benchmarks indicates a significant potential impact.
4. 8.9 Monte Carlo Tree Search with Boltzmann Exploration
   • Authors: Michael Painter, Mohamed Baioumy, Nick Hawes, Bruno Lacerda
   • Reason: The paper addresses a notable limitation in Monte-Carlo Tree Search by introducing new algorithms that utilize Boltzmann policies, which is expected to have a broad impact on automated planning and potentially other areas of reinforcement learning.
5. 8.7 Deep Reinforcement Learning Based Toolpath Generation for Thermal Uniformity in Laser Powder Bed Fusion Process
   • Authors: Mian Qin, Junhao Ding, Shuo Qu, Xu Song, Charlie C. L. Wang, Wei-Hsin Liao
   • Reason: The paper presents an innovative application of DRL for the manufacturing domain, which is uncommon, but with increasing industrial interest, it could become highly influential in the relevant fields.
6. 8.7 On the Sample Efficiency of Abstractions and Potential-Based Reward Shaping in Reinforcement Learning
   • Authors: Giuseppe Canonaco, Leo Ardon, Alberto Pozanco, Daniel Borrajo
   • Reason: This work tackles a crucial issue in RL, namely sample inefficiency, by using abstractions to automatically produce effective potential functions, which could be particularly impactful in the development of more efficient RL algorithms.
7. 8.5 Differentially Private Reinforcement Learning with Self-Play
   • Authors: Dan Qiao, Yu-Xiang Wang
   • Reason: Tackling privacy in multi-agent reinforcement learning is a timely and challenging issue. The paper’s exploration of differential privacy in the context of multi-agent systems could be impactful given the growing concerns over data privacy.
8. 8.5 Reinforcement Learning with Generalizable Gaussian Splatting
   • Authors: Jiaxu Wang, Qiang Zhang, Jingkai Sun, Jiahang Cao, Yecheng Shao, Renjing Xu
   • Reason: The paper presents a novel framework for representation in RL tasks, which could significantly influence vision-based RL through improved performance outcomes demonstrated in the RoboMimic environment.
9. 8.3 Medical mT5: An Open-Source Multilingual Text-to-Text LLM for The Medical Domain
   • Authors: Iker García-Ferrero, Rodrigo Agerri, Aitziber Atutxa Salazar, Elena Cabrio, Iker de la Iglesia, Alberto Lavelli, Bernardo Magnini, Benjamin Molinet, Johana Ramirez-Romero, German Rigau, Jose Maria Villa-Gonzalez, Serena Villata, Andrea Zaninello
   • Reason: The development of an open-source multilingual model for the medical domain addresses a gap in language technology for medical applications and could lead to advancements in natural language understanding and generation within healthcare AI.
10. 8.1 PINNACLE: PINN Adaptive ColLocation and Experimental points selection
    • Authors: Gregory Kang Ruey Lau, Apivich Hemachandra, See-Kiong Ng, Bryan Kian Hsiang Low
    • Reason: This paper introduces the first algorithm for jointly optimizing the selection of all training point types in Physics-Informed Neural Networks, potentially improving their generalization error and performance in various learning problems, which might have a substantial influence on PINN-based methods.