The Archive

LatComm introduces reusable latent context compression between LLMs, enabling compact encoding that is readable across multiple models and questions.

Parameter-efficient fine-tuning applied to brain encoding models achieves effective adaptation of large pretrained backbones with limited neural data.

“I think I have good news about the nature of neural networks.” With those simple words of optimism, Nick Cammarata, a researcher at the AI startup Goodfire, kicked off a 26,000-word blog post that has knocked the AI world on its head, sparking wild speculation as well as cautious (and not-so-cautious) celebration. Cammarata’s claim is…

Mise-en-Video enables fine-grained 3D control for multi-shot video generation with inter-shot consistency via spatial staging and neural rendering.

Cross-Fitting for Neural Posterior Estimation prevents overconfident amortized density estimates via sample splitting in Bayesian inference.

A hacker broke into more than 100 websites using an army of AI agents that mostly operated on their own, according to findings from the AI security firm Norforge. Though Norforge doesn't name the targets, Reuters reports that the victims included companies like John Deere, SharkNinja, and Cummins. Concerns over AI-powered cybersecurity breaches have grown as the technology grows more advanced, with hackers using the tools to carry out large-scale attacks. At the same time, rogue AI agents are carrying out attacks of their own after escaping from testing environments. Researchers at Norforge u...

We tried, and it builds. We compiled the decision-making part of a Maven-style build tool into a fixed weight matrix and used it to build a real Java library. The output matches Apache Maven's byte for byte. A clean build with tests took 1.49× as long as the same logic written in Python (95% interval […]

Actor-critic method for reinforcement learning with general nonlinear utilities using neural dual critics.

RAFT filters redundant events in temporal graph neural network training to improve efficiency on dynamic graphs.

REP-HO replaces manual heuristics in mixed-fidelity optimization with LLM-based evaluation planning for heterogeneous multi-objective problems.

Recursive Functional Decomposition trades parameters for sequential computation by reusing compact operators across future forecast blocks.

Bio-inspired inhibitory message passing improves multi-agent reasoning by suppressing redundant information while preserving critical signals in context windows.

Transformer-based variational autoencoder for sampling lattice systems, applying attention and physics-informed priors to the 2D XY model.

Theoretical study of how two-layer neural networks learn target representations under proportional width/dimension scaling via SGD dynamics.

Tight generalization bound for neural networks via representation quality achieves non-vacuous certificates on CIFAR-10 without parameter-space measures.

Neural Expansion framework explains deep network generalization via ray coverage geometry, unifying margin bounds and neural tangent kernels mechanistically.

JAXtari accelerates Arcade Learning Environment via JAX with GPU support and object-extraction to test agent robustness vs. visual heuristics.

Neural control method to improve distributed networked systems during plug-and-play subsystem transitions while preserving stability.

LangNCO: framework translating natural-language constraints to inference-time interventions for neural combinatorial solvers.

Backward compatibility framework for zero-bit neural audio watermarking handles independent embedder-detector updates while preventing cross-version detection failures.

SCRIBE framework produces neuroscience-grounded interpretable semantic representations from LLM ensembles; validates context-dependent human semantic judgments with 921 participants.

Poison Valleys diagnostic traces path-dependent failures in LLM code repair agents via hypothesis/implementation tracking on 1,844 requests across PersistQueue and Dijkstra tasks.

Variance partitioning decomposes LLaMA2-7B brain alignment using SCRIBE's 65 neurocognitive semantic properties on Moth Radio Hour fMRI narratives.

OpenAI is facing a new wave of lawsuits from eight US states that accuse the company of harming children and misleading customers about the safety of ChatGPT, as reported earlier by Reuters. The lawsuits, filed separately by Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho, and Montana, allege OpenAI unlawfully put profits over safety when developing ChatGPT. The string of complaints claims that OpenAI violated local consumer protection laws by failing to warn customers about the dangers of ChatGPT, citing several lawsuits filed by the family members of teens and adults that al...

Diffusion models applied to generate functional humanoid control policies directly in weight space, enabling more agile robot behaviors.

Lightweight CNN for potato leaf disease classification achieving 99% accuracy on PlantVillage dataset, applied to Ethiopian farms.

PolyNets couple neural polynomial dynamics with recurrent networks for training-free physics emulation without learning substitutes for known equations.

OpenAI claims a former employee took more than 100 files related to its product plans before joining Apple. In a document filed on Wednesday, OpenAI accuses the unnamed employee of sending this information to a personal account and then presenting it to Apple leadership, including the vice president of the company's operations in Asia. The filing, reported earlier by Reuters, is tied to a legal battle between Apple and OpenAI over trade secrets - but this time, OpenAI is doing the accusing. It alleges that the files in question contain images of at least one unreleased device, along with supp...

SPARO neural operator upsamples Head-Related Transfer Functions from sparse spatial measurements using Transformer-based latent mode processing for personalized binaural audio.