Seeing is Sorting: AI-Powered Textile Recognition for Automation
A new approach leverages digital twins and advanced visual AI to dramatically improve the accuracy and efficiency of automated textile sorting systems.
Science
Seeing, Understanding, and Acting: Guiding AI with Internal Signals
New research demonstrates a method for directly influencing the behavior of vision-language-action models by observing and manipulating their internal representations.
How Transformers ‘Think’ by the Book
![Deep linear neural networks and GPT-2, when trained on orthogonal data, exhibit feature cosine similarity patterns indicative of learned representations that distinguish between data sharing the same labels, suggesting an inherent capacity for discerning relationships beyond simple feature overlap-a characteristic quantifiable by [latex] \cos(\theta) [/latex] between feature vectors.](https://arxiv.org/html/2603.05143v1/2603.05143v1/Figure/linear_multi.png)
New research illuminates the mechanisms behind analogical reasoning in transformer models, revealing how they connect disparate concepts.
Robots That Never Forget: Continuous Learning Through Language
Researchers have developed a new framework that allows robots to continuously acquire and refine manipulation skills guided by natural language, without succumbing to catastrophic forgetting.
The Rise of Conversational AI: From Static Responses to Dynamic Interactions
A new wave of large language models is enabling real-time, interactive experiences, moving beyond simple text generation to truly dynamic conversations.
Robots Learn by Example: Building Skills from Human Feedback
A new method empowers robots to acquire and generalize skills by dynamically incorporating human guidance and successful code examples into a reusable knowledge base.
Coding Companions: Building AI Agents for the Command Line
Researchers detail the creation of an open-source AI assistant designed to tackle complex software engineering tasks directly within the terminal.
Staged Learning: Boosting Robot Skills with Smart Rewards
![Reward curricula demonstrably enhance the performance of both TD3 and SAC reinforcement learning algorithms across diverse robotic control tasks-DM Control, MobileRobot, and ManiSkill3-with optimal target weights of [latex]w_{\text{target}}=0.5[/latex], [latex]w_{\text{target}}=0.25[/latex], and a range of [latex]w_{\text{target}} \in \{0.25, 0.5, 0.75\}[/latex] respectively, as evidenced by consistently improved base and average rewards measured over the final 50,000 training steps and three random seeds.](https://arxiv.org/html/2603.05113v1/2603.05113v1/fig_ijcai/overview.png)
A new approach to reinforcement learning breaks down complex robotic tasks into manageable stages, improving training efficiency and adaptability.
Mapping Molecular Interactions with Graph Networks
A new deep learning approach harnesses the power of graph neural networks to predict how strongly proteins bind to other molecules.
Can AI Police Itself? Fact-Checking with Internal Knowledge
New research demonstrates that large language models can verify claims using the information already encoded within their parameters, eliminating the need for external databases.