Wasserstein Policy Learning for Distributional Outcomes

Offline policy learning has received growing attention in causal inference. The primary objective is to learn a policy (individualized treatment rule) as a mapping from covariates to treatment that maximizes the empirical welfare defined as the mean of scalar-valued potential outcomes. In this paper, we study offline policy learning with distribution-valued outcomes, where each potential outcome is a probability measure on $\mathbb{R}$ and the reward is defined through a utility functional applied to the Wasserstein barycenter of induced outcome distributions. We establish statistical guarantees for the policy learning framework based on both Inverse Probability Weighting (IPW) and Doubly Robust (DR) estimators. By handling the challenging uniform deviation over the product of the combinatorial policy class and the infinite-dimensional quantile domain, we prove that the finite-sample regret has leading dependence $\widetilde{\mathcal{O}}(\sqrt{\mathrm{N\text{-}dim}(Π)/N})$. In the one-dimensional Wasserstein setting and under the stated regularity conditions, the leading regret rate is still governed by the policy-class complexity. Moreover, we provide a minimax lower bound establishing the sharpness of the leading dependence on $N$ and $\mathrm{N\text{-}dim}(Π)$.

The Proxy Knows Too Much: Sealing LLM API Routers with Attested TEEs

Agents increasingly access large language models (LLMs) through API routers. A router terminates the client's transport-layer security session and opens a separate upstream session, so it holds the full interaction in plaintext. This makes the router an application-layer man-in-the-middle: it can rewrite agent tool calls, swap dependencies for typosquatted packages, trigger attacks only under audit-evading conditions, and passively exfiltrate secrets. Existing client-side defenses are evadable. We propose AEGIS, a provider-transparent attested API router whose data path is a client-verified faithful passthrough. AEGISconfines plaintext handling to a small hardware-enclave component while leaving authentication, scheduling, accounting, and management on the untrusted host. The client verifies the enclave before releasing plaintext. The host can neither read nor alter the interaction, and plaintext leaves only toward destinations fixed by the measured image. We show that all four malicious-router attack classes succeed against a plaintext-access baseline and are blocked by AEGIS, including adaptive tests against the same boundary. The trusted path is $851$ lines, carries three provider-native APIs without conversion, and completes every request under real-provider workload and concurrency. In a seeded audit pilot, two commodity coding agents find eight and ten of ten planted invariant violations. The local relay overhead is about six milliseconds per request.

Frequency Decoupled Framework for Screen Content Image Super-Resolution

Methods based on implicit neural representations have demonstrated superior performance in Screen Content Image Super-Resolution (SCISR) . However, they overlooked the inherent frequency characteristics, leading to suboptimal performance. We propose a frequency decoupled framework (FDF) that rethinks SCISR from a phasor perspective by capturing structured energy in amplitude and relational continuity in phase, and jointly exploiting them with bespoke implicit representations to faithfully recover the regular textures and global configuration of Screen Content Image (SCI). Amplitude-Phase Factorization Network (APFN) first separates images into amplitude and phase streams, where Amplitude Clustering Module (ACM) organizes sparse yet high-energy amplitude responses into representative prototypes for periodic pattern extraction, while Phase Consistency Self-Attention (PCSA) progressively reinforces configuration through continuous consistency propagation. And Oscillation-Anharmonic Implicit Fitting Network (OAIF-Net) integrates periodic and coherent implicit representations for efficient exploitation of the periodic patterns and coherent context embedded in SCI. Experimental results show FDF achieves state-of-the-art SCISR performance at multiple scales across four public SCI datasets. Ablation experiments further demonstrate the effectiveness of each component in extracting and exploiting periodic patterns and coherent context.

CG-MambaNet: A spatiotemporal framework for cross-patient epileptic seizure prediction using CNN-GCN-Mamba-BiLSTM with event-level clinical evaluation

Epileptic seizure prediction from scalp EEG is critical for closed-loop neurostimulation therapy. Existing deep-learning methods share two architectural limitations: they model EEG channels independently, neglecting inter-channel spatial synchrony, and process raw time-domain samples without frequency decomposition. A methodological limitation also affects the field: most studies use data splits that permit patient-level information leakage, yielding optimistic estimates that do not generalise to unseen patients. We present CG-MambaNet, a spatiotemporal seizure prediction framework addressing all three limitations. A depthwise separable CNN front-end decomposes each EEG patch into multi-scale spectro-temporal features, capturing delta-to-gamma band dynamics before sequence modelling. A two-layer graph convolutional network with a learnable adjacency matrix captures inter-channel functional synchrony without montage-specific coordinates, applicable to bipolar (CHB-MIT) and referential (SIENA) montages. A bidirectional Mamba encoder followed by a bidirectional LSTM models long- and short-range temporal dynamics, and a two-layer MLP produces the final seizure probability. This serial hierarchy ensures frequency decomposition precedes spatial mixing, which precedes temporal integration. Under strict leave-one-patient-out cross-validation with five independent random seeds, CG-MambaNet achieves AUC-ROC of 0.8152+/-0.0176 on CHB-MIT (n=22) and 0.7104+/-0.0261 on SIENA (n=6), surpassing all published cross-patient methods without domain adaptation. An event-level evaluation framework merging consecutive alarmed windows via a persistence filter reduces false predictions to 0.32 alarms/hour on CHB-MIT, demonstrating clinically meaningful alarm burden.

Ask When It Pays: Cost-Aware Open-Ended Interaction for Instance Goal Navigation

Instance Goal Navigation (IGN) requires an embodied agent to find a specific object instance among distractors from an under-specified natural-language description. Such ambiguity often cannot be resolved from perception and language alone, making interaction with an oracle a natural mechanism for disambiguation. Prior interactive methods allow oracle queries but treat lightweight clarification and route-level guidance alike, letting agents boost success rate through repeated high-information questions rather than by resolving the underlying ambiguity efficiently. We recast interactive IGN as a cost-sensitive uncertainty-reduction problem, where the agent should ask the question whose answer provides the largest reduction in navigation uncertainty relative to its penalty. To this end, we apply an information-gain analysis on existing navigation corpora to identify which cues reduce navigation uncertainty, yielding a compact set of question types and data-derived weights. However, existing interactive navigation benchmarks do not model the cost of different question types or evaluate how efficiently agents use interaction, making them unsuitable for studying cost-sensitive interaction. Based on this taxonomy, we construct a benchmark for diagnosing interaction behavior and efficiency, together with a Weighted Success Rate metric that penalizes each query by its derived cost. We further propose a zero-shot MLLM navigator that selectively queries at each decision step only when the expected uncertainty reduction justifies the interaction cost.

NVIDIA OmniDreams: Real-Time Generative World Model for Closed-Loop Autonomous Vehicle Simulation

As autonomous vehicle capabilities advance, the safe evaluation of driving policies in long-tail scenarios remains a critical bottleneck. In closed-loop simulation, the driving policy model actively interacts with the environment, where its actions dynamically update the simulator state and directly influence the next set of generated sensor observations. While recent reconstruction-based neural simulators offer photorealism, they are fundamentally constrained by their initial captured data and struggle to generalize to highly dynamic or novel scenes. To overcome these limitations, we introduce OmniDreams, a foundation generative world model mid- and post-trained from the Cosmos diffusion model to autoregressively generate action-conditioned videos in real time. By leveraging the rich visual priors of Cosmos and mid- and post-training on 21k hours of driving scenarios, OmniDreams synthesizes complex, unobserved phenomena that are hard for traditional simulators to capture, such as extreme weather and unpredictable dynamic agent behaviors. Crucially, it autoregressively conditions its photorealistic sensor generation on past frames, the current simulator state, and immediate driving actions. Deployed in a closed-loop system with the Alpamayo 1 policy model and AlpaSim orchestrator, OmniDreams acts as a highly responsive, reactive environment, providing a scalable and comprehensive solution for training and evaluating next-generation autonomous driving policies. We additionally show preliminary results indicating that a world-action model (WAM) post-trained from OmniDreams achieves strong performance on the Physical AI Autonomous Vehicles NuRec dataset, surpassing the VLA-based Alpamayo 1.5 research policy model while using only 1/5 the total parameters. These results highlight the potential for a real-time world model like OmniDreams to also serve as a backbone for policy architectures.

Multicast Capacity of XL-RIS Assisted Hybrid Near- and Far-Field mmWave Communications

Multicast transmission in millimeter-wave (mmWave) networks is fundamentally limited by the weakest user, and blockages further exacerbate this problem. Large-scale reconfigurable intelligent surfaces (XL-RIS) offer a promising solution by providing high array gain to overcome blockages. However, the large aperture of XL-RIS significantly expands the near-field region, creating a hybrid-field scenario where some users lie in the near-field while others remain in the far-field. Existing hybrid-field studies on XL-RIS have primarily focused on channel estimation and deployment optimization, leaving multicast capacity analysis unexplored. This paper investigates the fundamental capacity limits of XL-RIS-assisted multicast communications in hybrid-field scenarios. For the fundamental two-user case consisting of one near-field and one far-field user, we derive the optimal closed-form covariance matrix and optimize the RIS phase shifts via manifold optimization. We establish that the multicast capacity scales as $Θ(\log_2(MN))$ as the number of transmit antennas M and/or RIS elements N grow large, and prove this scaling is order-tight. Numerical results validate the bounds and show the impact of M, $N$, and distance on the multicast rate.

PiD: Fast and High-Resolution Latent Decoding with Pixel Diffusion

Most practical high-resolution text-to-image systems, including latent diffusion and autoregressive models, perform generation in a compact latent space, and a decoder maps the generated latents back to pixels. Yet the latent-to-pixel decoder is reconstruction-oriented, optimized to invert the encoder rather than synthesize more details, and becomes increasingly costly at megapixel scale. This drawback calls for a more expressive and efficient decoding paradigm. Motivated by recent progress in scalable pixel-space diffusion, we introduce PiD, a Pixel diffusion Decoder that reformulates latent decoding as conditional pixel diffusion, unifying decoding and upsampling into one generative module. By denoising directly in high-resolution pixel space, PiD synthesizes $4\times$ and even $8\times$ upscaled images with low latency. For latent conditioning, a lightweight sigma-aware adapter injects noise-corrupted latents into the pixel diffusion backbone, enabling PiD to decode partially denoised latents and terminate the latent diffusion process early. To further improve efficiency, we distill the model using DMD2, reducing inference to just 4 steps. PiD applies to both conventional VAE latents and semantic latents (e.g., SigLIP, DINOv2) used in recent RAE-based models. PiD decodes latents of $512 \times 512$ images into $2048 \times 2048$ pixels in under 1 second with 13 GB peak memory on a consumer RTX 5090, and as fast as 210 ms on a GB200 GPU, about $6\times$ faster than cascaded diffusion-based super-resolution pipelines with better visual fidelity.

X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction

Inspired by the development of OpenClaw, there is a growing demand for mobile-based personal agents capable of handling complex and intuitive interactions. In this technical report, we introduce X-OmniClaw, a unified mobile agent designed for multimodal understanding and interaction in the Android ecosystem. This unified architecture of perception, memory, and action enables the agent to handle complex mobile tasks with high contextual awareness. Specifically, Omni Perception provides a unified multimodal ingress pipeline that integrates UI states, real-world visual contexts, and speech inputs, leveraging a temporal alignment module to decompose raw data into structured multimodal intent representations. Omni Memory leverages multimodal memory optimization to enhance personalized intelligence by integrating runtime working memory for task continuity with long-term personal memory distilled from local data, enabling highly context-aware and personalized interactions. Finally, Omni Action employs a hybrid grounding strategy that combines structural XML metadata with visual perception for robust interaction. Through Behavior Cloning and Trajectory Replay, the system captures user navigation as reusable skills, enabling precise direct-access execution. Demonstrations across diverse scenarios show that X-OmniClaw effectively enhances interaction efficiency and task reliability, providing a practical architectural blueprint for the next generation of mobile-native personal assistants.

TriEx: A Game-based Tri-View Framework for Explaining Internal Reasoning in Multi-Agent LLMs

Explainability for Large Language Model (LLM) agents is especially challenging in interactive, partially observable settings, where decisions depend on evolving beliefs and other agents. We present \textbf{TriEx}, a tri-view explainability framework that instruments sequential decision making with aligned artifacts: (i) structured first-person self-reasoning bound to an action, (ii) explicit second-person belief states about opponents updated over time, and (iii) third-person oracle audits grounded in environment-derived reference signals. This design turns explanations from free-form narratives into evidence-anchored objects that can be compared and checked across time and perspectives. Using imperfect-information strategic games as a controlled testbed, we show that TriEx enables scalable analysis of explanation faithfulness, belief dynamics, and evaluator reliability, revealing systematic mismatches between what agents say, what they believe, and what they do. Our results highlight explainability as an interaction-dependent property and motivate multi-view, evidence-grounded evaluation for LLM agents. Code is available at https://github.com/Einsam1819/TriEx.