The Archive

MemHarness reconstructs retrieved memories contextually rather than replaying them verbatim, reducing negative transfer in LLM agents.

Locksmith Loop uses agentic test synthesis to validate COBOL-to-Java migrations via iterative branch penetration and witness search.

Theia auto-captions and validates Incidents1M, a multimodal disaster dataset for vision-language models via data-free knowledge distillation.

LLM-guided evolutionary search reformulates constraint models to improve solver efficiency on combinatorial problems.

CoMem exploits transformer depth division of labor to cache residuals at intermediate layers for unbounded-context memory on Qwen3-8B.

Topoformer encodes graph topology into attention-friendly sequences via Topo-Scan, enabling scalable graph representation learning.

OPAL improves online 3D bin packing by integrating operational constraints into placement-aware learning instead of geometry-only representations.

Survey of uncertainty quantification methods in deep learning, covering ensemble and Bayesian approaches for safety-critical deployment.

EgoGenesis synthesizes egocentric manipulation videos using geometry-aware conditioning and projective memory to expand embodied AI training data.

Opinion piece positioning Agent-as-a-Service (AaaS) as emerging paradigm for encapsulating agentic AI capabilities in virtual environments.

Empirical evaluation of 6 leading LLMs on Islamic knowledge tasks reveals hallucination and authenticity gaps in religious contexts.

CDAE method improves BERT robustness to textual perturbations via contrastive denoising autoencoder training.

EMBL AI Librarian provides structured knowledge layer for AI agents querying 40M+ life-sciences records from Europe PMC.

The Disrupt Stage is where many of the biggest conversations in technology happen, with a legacy that stretches back for more than a decade.

Qwen-UI-Agent technical report describes cross-platform GUI agent for mobile, web, CLI, and long-horizon task automation.

Survey of security threats across world-model lifecycle in embodied AI: data, sensors, state grounding, and physical execution.

Benchmark of 6 deep learning weather emulators (Pangu, FuXi, GraphCast, Aurora) on extreme heat forecasting at 10-15 day lead times.

Inverted causal self-attention mechanism for discovering causal structures in high-dimensional multivariate time series.

Vibe-FDTR agent framework enables LLMs to automate reproducible thermal property analysis from natural language requests.

Compressed LLMs pass quality guards but hallucinate procedure steps in agentic execution, exposing fidelity-safety gap.

Privacy-preserving federated learning framework for clinical EEG using masking-based secure aggregation and threshold secret sharing.

LLM pipeline for automating MADRS depression assessment scoring in clinical trial structured interviews.

Black-box evaluation reveals commercial multimodal content moderation APIs vulnerable to simple image transformations.

Gaussian noise injection prevents oversmoothing in deep recurrent GNNs by preserving representation diversity.

ReAlloc framework uses causal inference for multi-channel marketing budget allocation in e-commerce platforms.

TPACK-guided empirical study examines multi-agent AI tool integration into software engineering requirements quality curriculum.

AgenticASR refines speech transcription via agent-based approach that revises text when context clarifies from later speech.

Unified algorithmic framework for optimal decision trees compares search strategies to improve scalability.

Candidate-aware decoding enables adaptive early-exit in diffusion language models by position-specific commitment decisions.

The Series A was led by Khosla Ventures, with participation from Allianz, Rebel Fund, Brick and Mortar Ventures’ Darren Bechtel, and Y Combinator’s Garry Tan.