About
Research & Development
Challenging Audio Security: Discrete Optimal Transport as a Black-Box Attack
TLDR: A new research paper demonstrates that Discrete Optimal Transport (DOT) is a highly effective black-box adversarial attack against modern audio anti-spoofing systems. The...
Research & Development
Unlocking Dexterity: Strategic Touch Sensor Placement Enhances Robotic Hand Manipulation
TLDR: This research explores how different placements of tactile sensors on anthropomorphic robotic hands impact their ability to manipulate objects. Using deep reinforcement learning,...
Research & Development
Enhancing Underwater Vision: A New Model for Realistic Turbid Environment Simulation
TLDR: This paper introduces an improved synthetic underwater image generation model that addresses the limitations of traditional models, especially in highly turbid environments. By...
Research & Development
Formalizing Advanced Algebraic Geometry: The Multi-Graded Proj Construction in Lean
TLDR: This research paper details the formalization of the multi-graded Proj construction in Lean4, extending Grothendieck's original N-graded version to arbitrary finitely generated abelian...
Research & Development
FlowRL: Balancing Rewards for More Diverse LLM Reasoning Paths
TLDR: FlowRL is a new reinforcement learning method for large language models (LLMs) that improves reasoning by matching the full reward distribution instead of...
Research & Development
Unpacking Prosocial AI: Why High Scores Don’t Always Mean Better Cooperation
TLDR: A new study applied "Measurement Layouts" to analyze AI agents in the Melting Pot contest, revealing that top-performing agents often excel in scenarios...