The Research Desk.

The Machine-Native Economy: How Digital Assets Connect Intelligence, Commerce, and Compute

Will Su, Robert Mitchnick, Jay Jacobs, William Helm

BlackRock

SlopShape: Detecting AI-Written Blog Posts Through Structural Features

Jochen Madler

Sitefire

AI models collapse when trained on recursively generated data

Ilia Shumailov, Zakhar Shumaylov, Yarin Gal, Ross Anderson, Yiren Zhao, Nicolas Papernot

University of Oxford, University of Cambridge

Large-scale online deanonymization with LLMs

Simon Lermen, Daniel Paleka, Joshua Swanson, Michael Aerni, Nicholas Carlini, Florian Tramèr

ETH Zurich, Anthropic

Self-Play Pretraining with Zero Data

Affiliation: Tel Aviv University, Affiliation: Stanford University, Tel Aviv University, Stanford University

Language Model "Shape"

LLM Agents Can Easily Tamper With Their Own Traces

Affiliation: ELLIS Institute Tübingen, Affiliation: Max Planck Institute for Intelligent Systems, Affiliation: Tübingen AI Center, Affiliation: University of Tübingen perfect-crime.ai, Affiliation: Max Planck Institute for Intelligent Systems Affiliation: Tübingen AI Center

Guiding End-to-End Driving Models with Endpoint-Constrained Trajectory Optimization

InternW0- Δ Δ: A World Action Model Bridging Predictive Dynamics and Actions with 20K+ Hours of Open Data

Shanghai AI Laboratory

FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders

USC PSI Lab, Brown University, Rice University, University of Aberdeen, University of Notre Dame

SAGE: Mitigating Long-Horizon Reasoning Biases via Topological Guidance

Xinyue Zeng, Jiawei Zhang, Yujun Yan, Dawei Zhou

Disaggregated Quantization: Specializing LLM Prefill and Decode

Andrei Panferov, Maximilian Kleinegger, Sweta Priyadarshi, Tijmen Blankevoort, Dan Alistarh

IST Austria Distributed Algorithms and Systems Lab

CodeGraph: Open-Taxonomy Knowledge Graph for Source Code with Wikidata Grounding

Federico Pennino, Andrea Gurioli, Stefano Zacchiroli, Maurizio Gabbrielli, Paolo Ferragina

MOPD-Router: Rethinking Teacher Routing in Multi-Teacher On-Policy Distillation

Tianze Xu, Yanzhao Zheng, Zhentao Zhang, Yuanqiang Yu, Chao Ma

SII - GAIR

LightMIS: Ultra-Lightweight Medical Image Segmentation Without a Stage-Wise Decoder

Andrei Arhire, Mihaela-Elena Breabăn, Radu Timofte