Beyond Book Clubs: Quantifying Character Importance in Fiction
A new computational framework moves beyond simple interaction counts to analyze the nuanced significance of characters within novels.
Science
Beyond Alignment: Giving Users a Voice in AI Values
A new approach proposes shifting the focus from optimizing AI alignment to actively co-constructing it through ongoing user participation.
Remembering the Query: AgentSM Boosts Database Access with Reasoning Reuse
A new agentic system, AgentSM, dramatically improves the accuracy and efficiency of translating natural language into database queries by intelligently leveraging past reasoning steps.
Beyond Objects: Teaching Machines to Understand *How* Things Happen
![Models readily latch onto superficial object cues as shortcuts during learning, sacrificing robust verb representation-a study using a ViT[10] trained on a verb-object subset of Sth-com[16] reveals that while object accuracy increases rapidly, verb accuracy plummets in unseen compositional settings, even dropping below chance, demonstrating a bias towards easily-identified objects over generalized verb understanding.](https://arxiv.org/html/2601.16211v1/x5.png)
New research tackles a core challenge in video understanding: ensuring AI infers actions based on temporal reasoning, not just the objects present in a scene.
The Hidden Order in AI’s Black Box
![The system’s foundational principle centers on restructuring inference to achieve a cohesive and adaptable framework, where alterations to one component necessitate a comprehensive understanding of the interconnected whole to maintain systemic integrity and predictable behavior - a concept akin to biological organisms where structure fundamentally governs function [latex] S = f(I, R) [/latex], indicating structure (S) as a function of inference (I) and restructuring (R).](https://arxiv.org/html/2601.15871v1/G11.png)
New research reveals that large neural networks self-organize during training, opening the door to more efficient and scalable artificial intelligence.
Quadruped Robots Conquer Complex Terrain with Learned Perception
Researchers have developed a new framework that allows four-legged robots to navigate challenging environments by intelligently perceiving their surroundings and planning stable footholds.
Who Wrote This? Trust and AI in Scientific Summaries
New research explores how readers perceive the quality and trustworthiness of research abstracts created with or by artificial intelligence.
Seeing is Connecting: Boosting Robot Actions with Visual Clues
A new method helps robots better understand images and language by restoring crucial spatial relationships between objects, leading to more effective manipulation.
Beyond the Hype: Building Trustworthy AI for Legal Expertise
New research highlights a critical shift in AI design, demonstrating that systems grounded in verified knowledge sources are far more reliable than those that simply generate text.
Beyond Followers: Mapping the Dynamics of Social Connection
![User behavior exhibits a continuous spectrum across varying degrees of reciprocity, demonstrated by a heatmap revealing smooth transitions in user properties-represented as median values across a [latex]10 \times 10[/latex] grid defined by inbound and outbound reciprocity ratios-and indicating a lack of discrete behavioral boundaries.](https://arxiv.org/html/2601.15623v1/x31.png)
New research reveals how analyzing reciprocal interactions-not just who follows whom-unlocks a deeper understanding of user behavior and content engagement on social media.