research note
This paper addresses the challenge of improving large language model (LLM) reasoning through on-policy self-distillation (SD) with weaker privileged information (PI)
research note
This paper addresses the computational challenges of systematically modeling nuclear pasta phases that form in the inner crust of neutron stars
research note
This paper investigates whether vision-language models (VLMs), pretrained on both visual and textual data, provide improved alignment with human neural and behavioral signals during natural reading…
research note
This paper addresses a critical gap in the evaluation of Emotional Support Dialogue Systems (ESDSes) — current benchmarks primarily rely on simulated seekers who are cooperative and disclose informa…
research note
This paper addresses the challenge of maintaining resilient integrations with third-party APIs that often exhibit outages, throttling, latency spikes, or quota exhaustion outside the control of the…
research note
The paper addresses a key challenge in large language model (LLM) generated HTML — while many generated pages render visually at first glance, they fail under interactive use cases such as scrolling…
research note
This paper addresses the growing challenge of detecting synthetic images generated by diverse generative models, an increasingly urgent problem due to the proliferation of deepfakes and AI-generate…
research note
This paper tackles the problem of sequentially orchestrating a set of heterogeneous agents under uncertainty, where agents differ in reliability, cost, and output behavior, and these characteristic…
research note
This paper introduces LiveK12Bench, a novel benchmark designed to rigorously evaluate the reasoning capabilities of large multimodal models (LMMs) on authentic high school-level examinations
research note
This paper presents an in-depth spectroscopic and morphological study of UDG 6, a gas-rich star-forming ultra-diffuse galaxy (UDG) member of the Hydra I cluster, using MUSE integral field data from…
research note
QSignAI addresses the integration of quantum randomness and artificial intelligence (AI) in a deployed social participation system, leveraging recent breakthroughs recognized by the 2024–2025 Nobel…
research note
This paper addresses the persistent security problem of shared library hijacking in Linux systems, including embedded Linux
