Tag
The author reflects on four papers from CHI 2026, particularly one examining how AI writing tools with LLM-based suggestions can reduce gender bias in hiring evaluations, discussing ethical implications and personal inspiration.
VisionHOPE introduces a visual backbone that operates as a self-modifying learning system, where memory and learning rules co-evolve during image processing, achieving competitive results on benchmarks like ImageNet-1K, COCO, and ADE20K.
DRM introduces a Diffusion Reward Model for reward modeling as conditional density estimation, which outperforms baselines and enhances RLHF training by capturing multimodal reward structures.
A paper on DYSCO, which learns latent spaces to capture underlying dynamics and recover governing equations from noisy observations, has been accepted at Neurips.
Citation-Network v1.2.0 is an open-source app that builds interactive citation networks from Zotero libraries, with integration for Obsidian notes and a GUI for exploration.
Nature published a paper detailing Paper2Agent, a virtual corresponding author developed by Stanford's James Zou group, which enables users to interact with and execute methods from research papers through AI.
The paper derives a uniform concentration bound for two-timescale actor-critic algorithms with function approximation in reinforcement learning, analyzing the actor parameter's behavior with high probability.
This paper analyzes prompts to ChatGPT over two years, showing a shift towards indirect and implicit directives with less politeness, suggesting users are adapting to LLMs' inferential capabilities.
This paper presents an exploratory ablation study of TALH, a hybrid language model combining MLA and SSM, showing that SSM integration is more critical for validation performance than MLA in the tested setup, with insights on memory usage and timing on consumer hardware.
This academic paper explores the provenance and admissibility of machine self-reports in AI systems, discussing implications for trust, transparency, and legal contexts.
This paper replicates a distributional-semantics extractive summarization method for Hindi and evaluates it on standard corpora, finding that sentence position is the only contributing feature and current Hindi benchmarks fail to incentivize advanced content selection.
The paper introduces Hyperspherical Semantic Trajectory Analysis (HSTA), an unsupervised methodology for tracking technological diffusion using semantic embeddings from academic preprints and patent documents.
A user announces that their paper has been accepted to NeurIPS with review scores of 5-4-4, and they expect a formal notification soon.
The paper 'Open AI in the wild' examines real-world applications and community representation of open AI research, based on feedback from a conference author.
LEANN is a new vector indexing method that reduces storage requirements by about 97% through graph simplification and real-time embedding computation, while maintaining retrieval recall close to full HNSW. It is suitable for use on laptops and won the Best Paper award at MLSys 2026.
This paper introduces a discrete diffusion model using variational autoregressive networks to parameterize normalized probability distributions, applied to Ising models for accurate thermodynamic computations and enhanced Monte Carlo sampling.
Introduces Uncheatable Eval, a dynamic benchmark using compression rates to evaluate language models and mitigate data contamination.
This paper introduces WebMRE, an offline benchmark for multimodal web agents, and studies the mutual reinforcement effect between guide sentences and actions, demonstrating performance improvements and providing causal analysis of the guide as a causal channel.
TimeEvo is a failure-driven self-evolution method for time series agents that diagnoses capability gaps, synthesizes tools, and improves accuracy across tasks and backbones.
UniDataAgent is an ontology-grounded agent system designed for enterprise question-to-report automation, improving efficiency and accuracy by separating semantic acquisition from online execution, as demonstrated in a deployment at China Unicom.