AI & Machine Learning

Knowledge distillation can be performed by injecting 'experience' into prompts rather than updating model weights.

Gaussian Joint Embeddings provide a probabilistic alternative to deterministic SSL, eliminating the need for architectural asymmetries to prevent collapse.

A unified L0-gating mechanism that enables comparable sparsification and pruning across graphs, text, and tabular data.

Batch-level query routing for LLMs allows for strict cost and capacity control that per-query methods cannot achieve.

Achieves high-fidelity LiDAR densification in just 156ms while strictly enforcing sensor physics to prevent 'ghost points'.

Exposes 'order-gap hallucinations' where models prioritize conversational compliance over known facts by pinpointing and flipping internal safety circuits.

Proves that high scores on visual spatial benchmarks are achieved through token-level search (BFS in prose) rather than genuine visual planning.

Identifies a 'stability asymmetry' signature where deceptive models maintain stable internal beliefs while producing fragile, unstable external responses under perturbation.

Challenges the 'filter-first' data paradigm by showing that training on uncurated data with quality-score labels outperforms training on high-quality filtered subsets.

Introduces a 'clone-robust' mechanism (YRWR) to prevent AI model producers from strategically gaming the rankings in crowd-sourced arenas like Chatbot Arena.

Enables vision models to learn online from human corrections at inference time, reducing redundant manual effort in video segmentation by up to 34%.

Formalizes the 'Observability Gap' to explain why coding agents plateau: humans can only provide feedback on visible outputs, while bugs reside in invisible execution states.

Provides a high-dimensional theoretical foundation for why two-phase optimizers like DiLoCo are mathematically superior to standard SGD in specific noise regimes.

Mathematically proves that multi-agent planning workflows are decision-theoretically dominated by a centralized Bayes decision maker, setting fundamental limits on agentic emergent behavior.

Provides a formal proof that any semantic memory system (including RAG and vector retrieval) is mathematically guaranteed to suffer from interference and forgetting.

Demonstrates that Liquid Neural Networks can outperform Diffusion Policies in imitation learning with half the parameters and nearly 2x faster inference.

Achieves a 45x reduction in video generation inference latency and 2.5x higher training throughput using an efficient solution-flow framework.

Introduces neural topology probing to identify causally influential 'hub neurons' in Vision-Language Models that govern cross-modal behavior.

Identifies that the distinct 'AI prose style' (specifically em dash overuse) is a surviving artifact of markdown-saturated training data leaking into unstructured output.

Releases ROSClaw, a model-agnostic executive layer that allows any foundation model to control any ROS 2 robot through standardized capability discovery and safety envelopes.

Releases ChartNet, a million-scale, high-quality multimodal dataset for chart understanding spanning 24 chart types and 1.5 million samples.

Enables zero-shot monocular metric depth estimation across any camera type (fisheye, 360, ERP) using a single unified model.

Proposes a new reinforcement learning policy compression method based on long-horizon state-space coverage instead of immediate action-matching.

Introduces MeteoCap-3B, a billion-scale meteorological dataset with expert captions and a spectral-aware diffusion model for weather time-series generation.

Reframes LLM-assisted research as a scientific forecasting problem, training models to generate proposals that align with future (held-out) research directions.

Identifies that standard Transformer attention matrices are fundamentally ill-conditioned and proposes a drop-in 'preconditioned' replacement.

GSR-GNN achieves 30x training speedups and 87% memory reduction for deep Graph Neural Networks on circuit graphs.

A fully open industrial-scale pretraining project releasing 8T tokens of processed data, a 3B model, and 200+ controlled pretraining ablations.

Enables precise, physically plausible control over light position, color, and intensity in single images without a 3D model.

Systematically demonstrates that 'easy-to-hard' curriculum learning provides no benefit for LLM deductive reasoning tasks.

IP-SAM allows the Segment Anything Model (SAM) to perform automatic, prompt-free segmentation by generating its own 'intrinsic prompts'.

Challenges the necessity of discrete action tokenizers in robotics by using a continuous, single-stage flow matching policy.

Moves autonomous driving from 'predict-then-plan' to an interleaved VLA model where future frames and ego-actions are generated step-by-step.

A non-Turing-complete DSL that compiles high-level LLM routing and agent policies directly into verified infrastructure artifacts like Kubernetes NetworkPolicies.

Introduces a marketplace infrastructure that rebrands AI agents from mere tools into peer participants in a verifiable production network.

Scales Maximum Entropy population synthesis from 20 to 50+ categorical attributes by replacing exact expectation sums with Persistent Contrastive Divergence.

Reveals that the tight architectural coupling of image generation and understanding in unified models creates a new class of reciprocal safety vulnerabilities.

Introduces a vision model testbed that aligns AI visual attention (scanpaths) with human gaze without sacrificing classification accuracy.

Shows that standard task-completion benchmarks fail to distinguish agent capabilities and proposes 'Working Memory Fidelity' as a more predictive metric.

The first self-supervised, domain-agnostic model for LiDAR ground segmentation, eliminating the need for per-sensor manual labeling.

A production-grade framework that converts LLM/RAG evaluation into a deployment decision workflow using Pareto frontiers and CI gates.

Collapses the standard vision backbone-plus-decoder architecture into a single early-fusion Transformer stack for both perception and task modeling.

Couples visual representations directly into the RL optimization process (RLVR) for vision-language models using a structured reward reweighting mechanism.

A unified framework for neural network recombination that achieves state-of-the-art fine-tuning with fewer than 200 parameters.

Enables Active Learning for tabular data without model retraining by iteratively optimizing the 'labeled context' of foundation models.

Harmful intent in LLMs can be detected geometrically even after safety 'refusal' mechanisms have been surgically removed.

For LLM-driven optimization, complex meta-heuristics like simulated annealing are unnecessary; simple greedy hill climbing is a superior default.

Mathematical proof that LayerNorm structurally reduces model complexity compared to RMSNorm due to its mean-centering geometry.

Proposes 'Amdahl’s Law for AI,' proving that human effort in AI-assisted work is bottlenecked by the fraction of 'novel' tasks rather than agent capability.

Lie Generator Networks enable linear system identification with guaranteed physical stability and dissipation by construction rather than through loss penalties.