The Archive

StratCBT dataset of 9,688 psychological counseling sessions with CBT strategies for training dynamic LLM counselors.

Analysis connects RL algorithmic collusion in repeated games to Simple Penal Codes theory, proposing intervention mechanisms.

LLM agent groups replaying human deliberation overstate consensus by 1.5x due to higher participation and uniformity.

LLM verbal confidence reaches 0.956 AUROC for correctness prediction on TriviaQA, substantially outperforming agreement-based signals.

Huawei is accelerating the launch of its next-generation Ascend 960DT AI chip as it pushes to compete with Nvidia and close China’s AI computing gap with the U.S.

Today, I’m talking with Mustafa Suleyman, the CEO of Microsoft AI. As you’re no doubt aware, the biggest story in tech right now is the spiraling debate about AI safety and regulation. It should come as no surprise that Mustafa has strong opinions on how AI should be built and regulated. Microsoft just published a 37-page statement called the “Humanist AI Code of Conduct,” which lays out the company’s principles around AI development and even its philosophy around really thorny issues like AI consciousness. If you’ll recall from his last appearance on the show, Mustafa thinks companies like A...

In 34 of the 37 surveyed countries, people are more likely to believe AI will lead to job losses over the next 20 years. | Image: Pew Pew Research has published a new global survey that sheds light on how people view AI, including its impact on jobs, life in general, and income inequality. The survey questioned 42,151 people across 37 countries from February 8th to May 13th - well ahead of recent apocalyptic warnings. A majority sees AI as a threat to human employment. In 34 of the 37 countries surveyed, people are more likely to believe AI will lead to job losses over the next 20 years rathe...

People can use these assistants to make restaurant reservations and cancel subscriptions

A new coalition that includes Google, Nvidia, Anthropic and Emerald AI wants to find 100 GW of grid capacity for new data centers.

Federated learning approach reduces communication overhead in sharded data parallelism for large-scale foundation model training on heterogeneous multi-GPU clusters.

Diffusion-GAN hybrid generates 3D geological microstructures from 2D images for materials science simulation.

Theoretical analysis of minimax regret bounds for online contract design with discontinuous agent payoffs.

COMPASS algorithm combines parallel sub-solvers and learning for ordered clustered routing at 100K scale with non-local dependency awareness.

Qualitative reasoning framework enables AI agents to make spatial inferences for educational game tutoring and player guidance.

STR-Agent uses LLM to translate natural-language QoS requests into adaptive routing decisions for dynamic LEO satellite networks.

Latent bridge matching model synthesizes cross-strength brain MRI translations without task-specific architectures.

Signal-theoretic ML framework detects deceptive recruitment ads as labor exploitation indicator using multimodal textual and structural features.

Four-stage translation pipeline converts natural-language traffic rules to Prolog using reasoning-guided intermediates for computational inference.

Theoretical bounds on reconstruction error in autoencoders with identical forward maps applied before and after masking.

Pure single-loop extragradient method achieves linear convergence for strongly convex–concave minimax optimization with fixed parameters.

NeuSOGA3D combines neural implicit methods with symbolic geometric reasoning for interpretable 3D reconstruction from point clouds.

Hybrid speculative decoding strategy dynamically switches between neural drafting and context copying to maximize LLM inference speedup.

Matrix generalization of Harvey-Lawson cone proves exact special Lagrangian manifold structure in deep learning geometry.

QUALS framework uses pattern quantization and learnability synchronization for balanced training data in zero-shot time series forecasting.

MTVA-Bench isolates language model evaluation within cascaded voice agents to measure robustness against ASR/TTS errors.

Task-oriented semantic feature transmission bypasses image reconstruction, transmitting compressed features for low-SNR satellite remote sensing.

Surface-based diffusion bridge synthesizes FDG-PET from MRI on cortical manifolds for dementia biomarker accessibility.

Cognitive offloading to AI can reduce opportunities to practice skills, creating risks of deskilling. However, it remains unclear how to prevent deskilling without restricting access to AI. Here, we design two interventions to reduce offloading decisions: (1) metacognitive feedback that makes the implications of offloading for users explicit, and (2) an effort-based reward that incentivizes less extensive LLM assistance. We test both in a preregistered online experiment ($N = 704$) with a 2$\times$2 design and a no-AI control. The task was to practice fraction arithmetic with an LLM-based ass...

Vision-language models (VLMs) are fragile under image corruption. We find that the wording of the question affects VLMs in two opposite ways. Verbose questions make VLMs substantially more robust---e.g., rephrasing "Is there a cat?" into "Please look carefully and answer: is there a cat?". Conversely, VLMs become more fragile under corruption when the question is semantically complex or finer-grained, e.g., "what colour is the cup left of the chair?" instead of "is there a cup?". Both effects stem from question-conditioned cross-modal attention, which induces a spectral filter over image patc...

LSTM+EKF hybrid for identifying time-varying frequencies in linear dynamical systems using vibration data.