Zuyi Li

Jingxi Chen, Yixiao Zhang, Xiaoye Qian et al.

Images can be viewed as layered compositions, foreground objects over background, with potential occlusions. This layered representation enables independent editing of elements, offering greater flexibility for content creation. Despite the progress in large generative models, decomposing a single image into layers remains challenging due to limited methods and data. We observe a strong connection between layer decomposition and in/outpainting tasks, and propose adapting a diffusion-based inpainting model for layer decomposition using lightweight finetuning. To further preserve detail in the latent space, we introduce a novel multi-modal context fusion module with linear attention complexity. Our model is trained purely on a synthetic dataset constructed from open-source assets and achieves superior performance in object removal and occlusion recovery, unlocking new possibilities in downstream editing and creative applications.

Jianing Bai, Ren Wang, Zuyi Li

The advances in deep learning (DL) techniques have the potential to deliver transformative technological breakthroughs to numerous complex tasks in modern power systems that suffer from increasing uncertainty and nonlinearity. However, the vulnerability of DL has yet to be thoroughly explored in power system tasks under various physical constraints. This work, for the first time, proposes a novel physics-constrained backdoor poisoning attack, which embeds the undetectable attack signal into the learned model and only performs the attack when it encounters the corresponding signal. The paper illustrates the proposed attack on the real-time fault line localization application. Furthermore, the simulation results on the 68-bus power system demonstrate that DL-based fault line localization methods are not robust to our proposed attack, indicating that backdoor poisoning attacks pose real threats to DL implementations in power systems. The proposed attack pipeline can be easily generalized to other power system tasks.

Hongxing Ye, Zuyi Li

The increasing penetration of renewable energy in recent years has led to more uncertainties in power systems. In order to maintain system reliability and security, electricity market operators need to keep certain reserves in the Security-Constrained Economic Dispatch (SCED) problems. A new concept, deliverable generation ramping reserve, is proposed in this paper. The prices of generation ramping reserves and generation capacity reserves are derived in the Affine Adjustable Robust Optimization framework. With the help of these prices, the valuable reserves can be identified among the available reserves. These prices provide crucial information on the values of reserve resources, which are critical for the long-term flexibility investment. The market equilibrium based on these prices is analyzed. Simulations on a 3-bus system and the IEEE 118-bus system are performed to illustrate the concept of ramping reserve price and capacity reserve price. The impacts of the reserve credit on market participants are discussed.

Rui Liu, Dian Yu, Tong Zheng et al.

Reinforcement learning with verifiable rewards (RLVR) improves reasoning in large language models (LLMs) but struggles with exploration, an issue that still persists for multimodal LLMs (MLLMs). Current methods treat the visual input as a fixed, deterministic condition, overlooking a critical source of ambiguity and struggling to build policies robust to plausible visual variations. We introduce $\textbf{VOGUE (Visual Uncertainty Guided Exploration)}$, a novel method that shifts exploration from the output (text) to the input (visual) space. By treating the image as a stochastic context, VOGUE quantifies the policy's sensitivity to visual perturbations using the symmetric KL divergence between a "raw" and "noisy" branch, creating a direct signal for uncertainty-aware exploration. This signal shapes the learning objective via an uncertainty-proportional bonus, which, combined with a token-entropy bonus and an annealed sampling schedule, effectively balances exploration and exploitation. Implemented within GRPO on two model scales (Qwen2.5-VL-3B/7B), VOGUE boosts pass@1 accuracy by an average of 2.6% on three visual math benchmarks and 3.7% on three general-domain reasoning benchmarks, while simultaneously increasing pass@4 performance and mitigating the exploration decay commonly observed in RL fine-tuning. Our work shows that grounding exploration in the inherent uncertainty of visual inputs is an effective strategy for improving multimodal reasoning.

Yicheng He, Chengsong Huang, Zongxia Li et al.

Reinforcement learning (RL) provides a principled framework for improving Vision-Language Models (VLMs) on complex reasoning tasks. However, existing RL approaches often rely on human-annotated labels or task-specific heuristics to define verifiable rewards, both of which are costly and difficult to scale. We introduce VisPlay, a self-evolving RL framework that enables VLMs to autonomously improve their reasoning abilities using large amounts of unlabeled image data. Starting from a single base VLM, VisPlay assigns the model into two interacting roles: an Image-Conditioned Questioner that formulates challenging yet answerable visual questions, and a Multimodal Reasoner that generates silver responses. These roles are jointly trained with Group Relative Policy Optimization (GRPO), which incorporates diversity and difficulty rewards to balance the complexity of generated questions with the quality of the silver answers. VisPlay scales efficiently across two model families. When trained on Qwen2.5-VL and MiMo-VL, VisPlay achieves consistent improvements in visual reasoning, compositional generalization, and hallucination reduction across eight benchmarks, including MM-Vet and MMMU, demonstrating a scalable path toward self-evolving multimodal intelligence. The project page is available at https://bruno686.github.io/VisPlay/

Hongxing Ye, Yinyin Ge, Mohammad Shahidehpour et al.

In Part II of this two-part paper, we analyze the marginal prices derived in Part I of this two-part paper within a robust optimization framework. The load and generation are priced at Locational Marginal Price (LMP) while the uncertainty and generation reserve are priced at Uncertainty Marginal Price(UMP). The Financial Transmission Right (FTR) underfunding is demonstrated when there is transmission reserve. A comparison between traditional reserve price and UMP is presented. We also discuss the incentives for market participants within the new market scheme.

Xuan Liu, Zhiyi Li, Zuyi Li

Power systems become more prone to cyber-attacks due to the high integration of information technologies. In this paper, we demonstrate that the outages of some lines can be masked by injecting false data into a set of measurements. The success of the topology attack can be guaranteed by making that: 1)the injected false data obeys KCL and KVL to avoid being detected by the bad data detection program in the state estimation; 2)the residual is increased such that the line outage cannot be detected by PMU data. A quadratic programming problem is set up to determine the optimal attack vector that can maximize the residual of the outaged line. The IEEE 39-bus system is used to demonstrate the masking scheme.