We develop Morphlux, a server-scale programmable photonic fabric to interconnect accelerators within servers. We show that augmenting state-of-the-art torus-based ML data-centers with Morphlux can improve the bandwidth of tenant compute allocations by up to 66%, reduce compute fragmentation by up to 70%, and minimize the blast radius of chip failures. We develop a novel end-to-end hardware prototype of Morphlux to demonstrate these performance benefits which translate to 1.72X improvement in training throughput of ML models. By rapidly programming the server-scale fabric in our hardware testbed, Morphlux can replace a failed accelerator chip with a healthy one in 1.2 seconds.

The rapid growth of online video content, especially on short video platforms, has created a growing demand for efficient video editing techniques that can condense long-form videos into concise and engaging clips. Existing automatic editing methods predominantly rely on textual cues from ASR transcripts and end-to-end segment selection, often neglecting the rich visual context and leading to incoherent outputs. In this paper, we propose a human-inspired automatic video editing framework (HIVE) that leverages multimodal narrative understanding to address these limitations. Our approach incorporates character extraction, dialogue analysis, and narrative summarization through multimodal large language models, enabling a holistic understanding of the video content. To further enhance coherence, we apply scene-level segmentation and decompose the editing process into three subtasks: highlight detection, opening/ending selection, and pruning of irrelevant content. To facilitate research in this area, we introduce DramaAD, a novel benchmark dataset comprising over 800 short drama episodes and 500 professionally edited advertisement clips. Experimental results demonstrate that our framework consistently outperforms existing baselines across both general and advertisement-oriented editing tasks, significantly narrowing the quality gap between automatic and human-edited videos.

Large language models (LLMs) have recently demonstrated remarkable progress in formal theorem proving. Yet their ability to serve as practical assistants for mathematicians, filling in missing steps within complex proofs, remains underexplored. We identify this challenge as the task of subgoal completion, where an LLM must discharge short but nontrivial proof obligations left unresolved in a human-provided sketch. To study this problem, we introduce FormalML, a Lean 4 benchmark built from foundational theories of machine learning. Using a translation tactic that converts procedural proofs into declarative form, we extract 4937 problems spanning optimization and probability inequalities, with varying levels of difficulty. FormalML is the first subgoal completion benchmark to combine premise retrieval and complex research-level contexts. Evaluation of state-of-the-art provers highlights persistent limitations in accuracy and efficiency, underscoring the need for more capable LLM-based theorem provers for effective subgoal completion,

Neural linguistic steganography aims to embed information into natural text while preserving statistical undetectability. A fundamental challenge in this ffeld stems from tokenization ambiguity in modern tokenizers, which can lead to catastrophic decoding failures. The recent method, SyncPool, addresses this ambiguity by employing a coarse-grained synchronization mechanism over groups of ambiguous candidates. However, SyncPool sacriffces embedding capacity, as it utilizes the entire Shannon entropy of an ambiguous group solely for synchronization rather than for payload embedding. We propose a method named look-ahead Sync, which overcomes the capacity limitation of SyncPool while retaining its provable security guarantees. Our approach performs minimal synchronized sampling only on truly indistinguishable token sequences, while strategically preserving all other discernible paths to maximize embedding capacity. We provide theoretical proofs for the security of our method and analyze the gap between its achievable embedding capacity and the theoretical upper bound. Experiments on English (using Llama 3) and Chinese (using Qwen 2.5) benchmarks show that our method consistently approaches the theoretical capacity upper bound and signiffcantly outperforms SyncPool. The improvement in embedding rate exceeds 160% in English and 25% in Chinese, particularly in settings with larger candidate pools. This work represents a signiffcant step toward practical high-capacity provably secure linguistic steganography.

Training long-context language models to capture long-range dependencies requires specialized data construction. Current approaches, such as generic text concatenation or heuristic-based variants, frequently fail to guarantee genuine long-range dependencies. We propose EntropyLong, a novel data construction method that leverages predictive uncertainty to verify dependency quality. Our approach identifies high-entropy positions in documents, retrieves semantically relevant contexts from large corpora, and verifies their utility by assessing whether they reduce prediction entropy. This model-in-the-loop verification ensures each dependency represents measurable information gain rather than spurious correlation. We construct training samples with long-range dependencies by combining original documents with these verified contextual supplements. Using FineWebEdu and Cosmopedia, we generate a dataset of 128K-length sequences with verified dependencies. Models trained on this data demonstrate significant improvements on RULER benchmarks, particularly in tasks requiring distant information. Following instruction fine-tuning, our models also achieve substantial gains on LongBenchv2, demonstrating enhanced long-context understanding. Extensive ablation studies further validate the necessity and effectiveness of entropybased verification for long-context training.

Empirical scaling laws establish a relationship between the number of parameters and model capability, as evidenced by models with hundreds of billions of parameters achieving state-of-the-art results on benchmarks (Kaplan et al., 2020; Grattafiori et al., 2024). However, the autoregressive generation process requires accessing all model parameters for each forward pass, creating a memory bandwidth bottleneck that impacts token generation latency. Furthermore, current trends toward more sophisticated LLM applications, such as multi-hop reasoning (Wei et al., 2022), tool integration (Patil et al., 2024), and reasoning capability (Y ang et al., 2025; DeepMind, 2025), produce longer output sequences, amplifying the computational burden of inference. One prominent approach to address inference latency is Speculative Decoding (SD), which achieves partial parallelization of the generation process (Leviathan et al., 2023; Chen et al., 2023). Standard SD operates by deploying a computationally efficient draft model to propose candidate token sequences, which are subsequently validated in parallel by the target model (the model of interest). The acceptance criterion for draft tokens relies on a probability-based alignment verification: draft tokens are accepted when their likelihood under the target model meets or exceeds their likelihood under the draft model.

Medical Multimodal Large Language Models (Med-MLLMs) have shown great promise in medical visual question answering (Med-VQA). However, when deployed in low-resource settings where abundant labeled data are unavailable, existing Med-MLLMs commonly fail due to their medical reasoning capability bottlenecks: (i) the intrinsic reasoning bottleneck that ignores the details from the medical image; (ii) the extrinsic reasoning bottleneck that fails to incorporate specialized medical knowledge. To address those limitations, we propose AMANDA, a training-free agentic framework that performs medical knowledge augmentation via LLM agents. Specifically, our intrinsic medical knowledge augmentation focuses on coarse-to-fine question decomposition for comprehensive diagnosis, while extrinsic medical knowledge augmentation grounds the reasoning process via biomedical knowledge graph retrieval. Extensive experiments across eight Med-VQA benchmarks demonstrate substantial improvements in both zero-shot and few-shot Med-VQA settings. The code is available at https://github.com/REAL-Lab-NU/AMANDA.

Their ability to continuously listen and respond to user queries opens new possibilities for human-machine interfaces in multiple applications, especially for those that require fast turnaround between humans and machines, such as question answering and brainstorming. Thanks to the recent developments of large transformers, these models can be implemented as direct speech-to-speech (S2S) models [3, 4, 5]. Because these S2S models are realized by a monolithic architecture that does not need to synchronize with other systems, their turnaround time is typically very low, contributing to more natural interaction. Moshi [2] is a pioneering model in that direction with an end-to-end S2S model for full-duplex conversational AI. However, because the inputs and outputs of S2S models are information-rich acoustic signals, auto-regressive modeling presents a fundamental challenge related to model capacity. In other words, unlike in text-only large language models (LLMs), S2S modeling must capture not only the verbal content of speech but also paralin-guistic features, such as speaking style, emotion, etc. This creates a fundamental inefficiency for S2S models with regard to knowledge acquisition. Given a similar model size, a text-only LLM can learn more knowledge because its capacity is dedicated solely to text, Figure 1.

Large language models accelerate literature synthesis but can hallucinate and mis-cite, limiting their usefulness in expert workflows. We present RA-FSM (Research Assistant - Finite State Machine), a modular GPT-based research assistant that wraps generation in a finite-state control loop: Relevance -> Confidence -> Knowledge. The system is grounded in vector retrieval and a deterministic citation pipeline. The controller filters out-of-scope queries, scores answerability, decomposes questions, and triggers retrieval only when needed, and emits answers with confidence labels and in-corpus, de-duplicated references. A ranked-tier ingestion workflow constructs a domain knowledge base from journals, conferences, indices, preprints, and patents, writing both to a dense vector index and to a relational store of normalized metrics. We implement the system for photonics and evaluate it on six task categories: analytical reasoning, numerical analysis, methodological critique, comparative synthesis, factual extraction, and application design. In blinded A/B reviews, domain experts prefer RA-FSM to both a strong Notebook LM (NLM) and a vanilla Default GPT API call single-pass baseline, citing stronger boundary-condition handling and more defensible evidence use. Coverage and novelty analyses indicate that RA-FSM explores beyond the NLM while incurring tunable latency and cost overheads. The design emphasizes transparent, well-cited answers for high-stakes technical work and is generalizable to other scientific domains.

Abstract--This paper introduces Agentic-AI Healthcare, a privacy-aware, multilingual, and explainable research prototype developed as a single-investigator project. The platform integrates a dedicated Privacy & Compliance Layer that applies role-based access control (RBAC), AES-GCM field-level encryption, and tamper-evident audit logging, aligning with major healthcare data protection standards such as HIPAA (US), PIPEDA (Canada), and PHIPA (Ontario). Example use cases demonstrate multilingual patient-doctor interaction (English, French, Arabic) and transparent diagnostic reasoning powered by large language models. As an applied AI contribution, this work highlights the feasibility of combining agentic orchestration, multilingual accessibility, and compliance-aware architecture in healthcare applications. This platform is presented as a research prototype and is not a certified medical device. This paper presents a working prototype that integrates agentic orchestration via the Model Context Protocol (MCP), field-level encryption, and multilingual LLM agents into a single compliance-aware stack for healthcare.