Electricity demands are increasing significantly and the traditional power grid system is facing huge challenges. As the desired next-generation power grid system, smart grid can provide secure and reliable power generation, and consumption, and can also realize the system’s coordinated and intelligent power distribution. Coordinating grid power distribution usually requires mutual communication between power distributors to accomplish coordination. However, the power network is complex, the network nodes are far apart, and the communication bandwidth is often expensive. Therefore, how to reduce the communication bandwidth in the cooperative power distribution process task is crucially important. One way to tackle this problem is to build mechanisms to selectively send out communications, which allow distributors to send information at certain moments and key states. The distributors in the power grid are modeled as reinforcement learning agents, and the communication bandwidth in the power grid can be reduced by optimizing the communication frequency between agents. Therefore, in this paper, we propose a model for deciding whether to communicate based on the causal inference method, Causal Inference Communication Model (CICM). CICM regards whether to communicate as a binary intervention variable, and determines which intervention is more effective by estimating the individual treatment effect (ITE). It offers the optimal communication strategy about whether to send information while ensuring task completion. This method effectively reduces the communication frequency between grid distributors, and at the same time maximizes the power distribution effect. In addition, we test the method in StarCraft II and 3D environment habitation experiments, which fully proves the effectiveness of the method.

Deep Research systems have revolutionized how LLMs solve complex questions through iterative reasoning and evidence gathering. However, current systems remain fundamentally constrained to textual web data, overlooking the vast knowledge embedded in multimodal documents: scientific papers, technical reports, and financial documents where critical information exists in figures, tables, charts, and equations. Processing such documents demands sophisticated parsing to preserve visual semantics, intelligent chunking to maintain structural coherence, and adaptive retrieval across modalities, which are capabilities absent in existing systems. In response, we present Doc-Researcher, a unified system that bridges this gap through three integrated components: (i) deep multimodal parsing that preserves layout structure and visual semantics while creating multi-granular representations from chunk to document level, (ii) systematic retrieval architecture supporting text-only, vision-only, and hybrid paradigms with dynamic granularity selection, and (iii) iterative multi-agent workflows that decompose complex queries, progressively accumulate evidence, and synthesize comprehensive answers across documents and modalities. To enable rigorous evaluation, we introduce M4DocBench, the first benchmark for Multi-modal, Multi-hop, Multi-document, and Multi-turn deep research. Featuring 158 expert-annotated questions with complete evidence chains across 304 documents, M4DocBench tests capabilities that existing benchmarks cannot assess. Experiments demonstrate that Doc-Researcher achieves 50.6% accuracy, 3.4× better than state-of-the-art baselines, validating that effective document research requires not just better retrieval, but fundamentally deep parsing that preserve multimodal integrity and support iterative research. Our work establishes a new paradigm for conducting deep research on multimodal document collections.

To train generalizable Reinforcement Learning (RL) agents, researchers recently proposed the Unsupervised Environment Design (UED) framework, in which a teacher agent creates a very large number of training environments and a student agent trains on the experiences in these environments to be robust against unseen testing scenarios. For example, to train a student to master the “stepping over stumps” task, the teacher will create numerous training environments with varying stump heights and shapes. In this paper, we argue that UED neglects training efficiency and its need for very large number of environments (henceforth referred to as infinite horizon training) makes it less suitable to training robots and non-expert humans. In real-world applications where either creating new training scenarios is expensive or training efficiency is of critical importance, we want to maximize both the learning efficiency and learning outcome of the student. To achieve efficient finite horizon training, we propose a novel Markov Decision Process (MDP) formulation for the teacher agent, referred to as Unsupervised Training Sequence Design (UTSD). Specifically, we encode salient information from the student policy (e.g., behaviors and learning progress) into the teacher's state space, enabling the teacher to closely track the student's learning progress and consequently discover the optimal training sequences with finite lengths. Additionally, we explore the teacher's efficient adaptation to unseen students at test time by employing the context-based meta-learning approach, which leverages the teacher's past experiences with various students. Finally, we empirically demonstrate our teacher's capability to design efficient and effective training sequences for students with varying capabilities.

Recent work on designing an appropriate distribution of environments has shown promise for training effective generally capable agents. Its success is partly because of a form of adaptive curriculum learning that generates environment instances (or levels) at the frontier of the agent's capabilities. However, such an environment design framework often struggles to find effective levels in challenging design spaces and requires costly interactions with the environment. In this paper, we aim to introduce diversity in the Unsupervised Environment Design (UED) framework. Specifically, we propose a task-agnostic method to identify observed/hidden states that are representative of a given level. The outcome of this method is then utilized to characterize the diversity between two levels, which as we show can be crucial to effective performance. In addition, to improve sampling efficiency, we incorporate the self-play technique that allows the environment generator to automatically generate environments that are of great benefit to the training agent. Quantitatively, our approach, Diversity-induced Environment Design via Self-Play (DivSP), shows compelling performance over existing methods.