Anchor-based 3D Gaussian Splatting (GS), exemplified by Scaffold-GS, achieves remarkable storage efficiency through a hybrid explicit-implicit representation. However, their reliance on a single, monolithic network to decode anchor features imposes a severe bottleneck on model capacity, often resulting in blurred details and view-dependent artifacts in complex scenes. To break this bottleneck, we introduce the concept of Scene Experts: a strategy that decomposes the task of modeling a complex scene across a collection of specialized sub-models. To realize the paradigm, we propose MoE-GS. Our approach designs the decoder as a Sparsely-Gated Mixture of Experts (MoE), which dramatically increases the model's total capacity while maintaining comparable inference cost via sparse activation. To effectively train this high-capacity model, we propose two key innovations: (1) A progressive curriculum learning strategy that first trains all experts on a robust baseline before encouraging them to specialize on different scene components. (2) A novel opacity-aware regularization that penalizes inactive neural Gaussians, ensuring the expanded capacity is efficiently used. Extensive experiments demonstrate that MoE-GS substantially outperforms state-of-the-art methods on diverse benchmarks, significantly improving reconstruction fidelity while requiring a smaller or comparable Gaussian model size.

The escalating energy consumption and carbon footprint of training large-scale Web AI models pose urgent challenges for sustainable development. Dataset Distillation (DD) offers a promising avenue for green AI by compressing large datasets into small synthetic ones for efficient training. However, most existing DD methods overfit to the inductive biases of specific source architectures (e.g., CNNs or ViTs), resulting in poor cross-model generalization. This limitation necessitates redundant re-distillation processes for different architectures, severely undermining the energy-saving potential of DD. To address this, we introduce Adversarial Mutual Information Distillation (AMID), a rigorous framework designed to create highly reusable and robust synthetic datasets. From an information-theoretic perspective, we cast model-agnosticism as minimizing the mutual information (MI) between the synthetic data and the specific identity of the distillation model. We convert this intractable objective into a tractable two-player adversarial game, which unifies knowledge preservation with adversarial unlearning of architectural bias. Extensive experiments on CIFAR-10 and Tiny ImageNet demonstrate that AMID achieves state-of-the-art cross-architecture generalization across diverse CNNs and ViTs. Crucially, our analysis confirms that AMID significantly reduces the computational overhead and CO2 emissions of downstream training while maintaining robust performance, paving the way for energy-efficient, transferable, and sustainable Web AI ecosystems.

The victim cache was originally designed as a secondary cache to handle misses in the L1 data (L1D) cache in CPUs. However, this design is often sub-optimal for GPUs. Accessing the high-latency L1D cache and its victim cache can lead to significant latency overhead, severely degrading the performance of certain applications. We introduce HIVE, a high-priority victim cache designed to accelerate GPU memory accesses. HIVE handles memory requests first, before they reach the L1D cache. Our experimental results show that HIVE achieves an average performance improvement of $\mathbf{7 7. 1 \%}$ and $\mathbf{2 1. 7 \%}$ compared to the baseline and the state-of-the-art architecture, respectively.

In recent years, new methods for solving partial differential equations (PDEs) such as Monte Carlo random walk methods have gained considerable attention. However, due to the lack of hardware-intrinsic randomness in the conventional von Neumann architecture, the performance of PDE solvers is limited. In this paper, we introduce NeuroPDE, a hardware design for neuromorphic PDE solvers that utilizes emerging spintronic and ferroelectric devices. NeuroPDE incorporates spin neurons that are capable of probabilistic transmission to emulate random walks, along with ferroelectric synapses that store continuous weights non-volatilely. The proposed NeuroPDE achieves a squared error of less than 1e-2 compared to analytical solutions when solving diffus3.48× to 315× speedup in execution time and an energy consumption advantage of 2.7× to 29.8× over advanced CMOS-based neuromorphic chips. By leveraging the inherent physical stochasticity of emerging devices, this study paves the way for future probabilistic neuromorphic computing systems.

Simulating society with large language models (LLMs), we argue, requires more than generating plausible behavior; it demands cognitively grounded reasoning that is structured, revisable, and traceable. LLM-based agents are increasingly used to emulate individual and group behavior, primarily through prompting and supervised fine-tuning. Yet current simulations remain grounded in a behaviorist “demographics in, behavior out” paradigm, focusing on surface-level plausibility. As a result, they often lack internal coherence, causal reasoning, and belief traceability—making them unreliable for modeling how people reason, deliberate, and respond to interventions.To address this, we present a conceptual modeling paradigm, Generative Minds (GenMinds), which draws from cognitive science to support structured belief representations in generative agents. To evaluate such agents, we introduce the RECAP (REconstructing CAusal Paths) framework, a benchmark designed to assess reasoning fidelity via causal traceability, demographic grounding, and intervention consistency. These contributions advance a broader shift: from surface-level mimicry to generative agents that simulate thought—not just language—for social simulations.

Federated Distillation (FD) is an innovative distributed machine learning paradigm that enables efficient and flexible cross-device knowledge transfer through knowledge distillation, without the need to upload large-scale model parameters to a central server. Although FD has attracted increasing attention in recent years, its security aspects remain relatively underexplored. Existing attack methods targeting traditional federated learning mainly focus on the transmission of model parameters and gradients, while attacks specifically designed for the unnormalized outputs (logits) in the emerging FD paradigm are still lacking. To fill this research gap and contribute to the enhancement of FD’s security, we previously proposed the Federated Distillation Logits Attack (FDLA), which manipulates the logits transmitted during communication to mislead and degrade the performance of client models. However, FDLA has limitations in controlling its impact on participants with different roles or identities and lacks a systematic investigation into the effects of malicious interventions at various stages of knowledge transfer. To overcome these limitations, we propose a more advanced and controllable logits poisoning method—Peak-Controlled Federated Distillation Logits Attack (PCFDLA). PCFDLA enhances the effectiveness of FDLA by precisely controlling the peak values of logits to adjust the intensity of the attack. This method generates highly misleading perturbations that achieve stronger attack performance while maintaining a similar level of stealthiness to FDLA when detection is based on differences in model parameters. Moreover, we introduce a novel evaluation metric to more comprehensively assess the performance of such attacks. Experimental results show that PCFDLA significantly increases the destructive impact on victim models while maintaining high stealth. It consistently achieves superior performance across multiple datasets, highlighting its potential threat to the security of federated distillation systems.

We introduce a learning framework for automated floorplan generation which combines generative modeling using deep neural networks and user-in-the-loop designs to enable human users to provide sparse design constraints. Such constraints are represented by a layout graph. The core component of our learning framework is a deep neural network, Graph2Plan, which converts a layout graph, along with a building boundary, into a floorplan that fulfills both the layout and boundary constraints. Given an input building boundary, we allow a user to specify room counts and other layout constraints, which are used to retrieve a set of floorplans, with their associated layout graphs, from a database. For each retrieved layout graph, along with the input boundary, Graph2Plan first generates a corresponding raster floorplan image, and then a refined set of boxes representing the rooms. Graph2Plan is trained on RPLAN, a large-scale dataset consisting of 80K annotated floorplans. The network is mainly based on convolutional processing over both the layout graph, via a graph neural network (GNN), and the input building boundary, as well as the raster floorplan images, via conventional image convolution. We demonstrate the quality and versatility of our floorplan generation framework in terms of its ability to cater to different user inputs. We conduct both qualitative and quantitative evaluations, ablation studies, and comparisons with state-of-the-art approaches.

It is believed that nonverbal visual information including facial expressions, facial micro-actions and head movements plays a significant role in fundamental social communication. Unfortunately it is regretful that the blind can not achieve such necessary information. Therefore, we propose a social assistant system, Lookine, to help them to go beyond this limitation. For Lookine, we apply the novel techniques including facial expression recognition, facial action recognition and head pose estimation, and obey barrier-free principles in our design. In experiments, the algorithm evaluation and user study prove that our system has promising accuracy, good real-time performance, and great user experience.