Articles

Every paper breakdown, newest first.

19 min read

Orca: What If Next-Token, Next-Frame, and Next-Action Are the Same Task?

Orca (BAAI) replaces next-token, next-frame, and next-action prediction with a single Next-State-Prediction objective. A frozen 4B backbone pre-trained on 12.5K hours of video — with zero action labels — feeds three lightweight readouts, and the action readout, trained on just 200 trajectories per task, beats π0.5 on OOD robot manipulation (32.4 vs 29.4).

world-modelsnext-state-predictionrepresentation-learning
15 min read

SWE-Explore: The Benchmark That Finally Asks — Did Your Coding Agent Read the Right Code?

SWE-Explore isolates repository exploration from patch generation, revealing that coding agents find the right files ~65% of the time but recall only ~15-19% of the lines that actually matter — and that context efficiency predicts downstream resolve rate with Pearson r = 0.950.

coding-agentsbenchmarkscode-localization
13 min read

SkillOpt: Training Agent Skills Like Neural Network Weights - Without Touching the Model

SkillOpt applies deep-learning-style optimization - bounded edit budgets, validation gating, rejected-edit memory - to natural-language skill documents, improving frozen LLMs by up to +39 points across 52/52 evaluated cells without changing a single model weight.

agentsprompt-optimizationllm
8 min read

MolmoAct2: The First Fully Open Robot Controller That Beats Closed-Source Giants

MolmoAct2 is a fully open vision-language-action model that outperforms π0.5 and matches Gemini Robotics ER, achieving 97.2% on LIBERO and 87.1% real-world success via per-layer KV-cache conditioning and adaptive depth reasoning.

roboticsVLAembodied-reasoning
11 min read

RecursiveMAS: What If Your Multi-Agent System Was Just One Big Recursive Neural Network?

RecursiveMAS treats an entire multi-agent system as a single recursive computation in latent space, adding only 0.31% trainable parameters while achieving +8.3% accuracy, 2.4x speedup, and 75.6% token reduction over text-based multi-agent baselines.

AImulti-agent-systemsrecursive-computation
12 min read

Tstars-Tryon 1.0: Virtual Try-On as Multi-Image Editing at Taobao Scale

How a unified 5B MMDiT trained with multi-reward RL and step distillation reframes virtual try-on as multi-image editing — and runs in under 4 seconds in production.

AIdiffusionvirtual-try-on
9 min read

ClawGUI: A Full-Stack Open-Source Pipeline for GUI Agents

ClawGUI unifies online RL training, reproducible evaluation, and real-device deployment of GUI agents into one open-source pipeline — and shows a 2B model trained inside it can beat 72B untrained baselines on MobileWorld.

gui-agentsreinforcement-learningbenchmarking
10 min read

SkillClaw: Making LLM Agent Skills Evolve Collectively

SkillClaw is a framework for collective skill evolution in multi-user LLM agent ecosystems. Instead of static skill libraries, the system automatically learns from interactions across users and propagates improvements to everyone.

llm-agentsskill-evolutionmulti-user
7 min read

TAPS: Why Your Draft Model's Training Data Matters More Than Its Architecture

TAPS shows draft training data drives speculative decoding: merged-tree verification of math and chat specialists hits 5.11 acceptance length, weight averaging 2.59.

AINLPLLM
5 min read

Demystifying Video Reasoning: Models Don't Think in Frames - They Think in Denoising Steps

Chain-of-Steps shows video diffusion models reason across denoising steps, not frames, exploring and self-correcting before converging on a single answer.

AIComputer VisionVideo Generation
4 min read

Seoul World Model: AI That Generates Video of Real Cities From Street Photos

Seoul World Model generates photorealistic city video from street-view photos, cutting FID on unseen Busan to 28.4 from 62.1 via cross-temporal reference pairing.

AIComputer VisionVideo Generation
4 min read

Lost in Stories: How LLMs Lose the Thread in Long Narratives

ConStory-Bench measures consistency in 10,000-word stories: GPT-5-Reasoning leads with 0.113 errors per 10k words, and its checker finds 3.2x more bugs than experts.

AILLMNLP
4 min read

Utonia: One Encoder For All Point Clouds

Utonia trains one 137M encoder for indoor, LiDAR, satellite, CAD and video point clouds. Stripped of color it scores 77.0% mIoU on ScanNet vs Concerto's 36.8%.

AI3DPoint Clouds
4 min read

SAGE: Your Reasoning Model Knows When to Stop Thinking — You Just Won't Let It

SAGE scores reasoning paths by cumulative log-probability instead of next-token probability, letting models stop thinking 40-50% earlier at equal or better accuracy.

AILLMReasoning
4 min read

When GPT Discovers Physics: A Breakthrough in Gluon Theory

GPT-5.2 Pro proposed the formula showing single-minus gluon amplitudes are nonzero in Klein space, valued only in -1, 0 or 1; it passed five consistency tests.

AIGPTPhysics
4 min read

OPUS: How to Train LLMs 6x Faster by Choosing the Right Data

OPUS scores every sample by alignment minus redundancy in optimizer update space; on Qwen3-8B it beat full 3B-token training using 0.5B, at 4.7% extra compute.

AILLMMachine Learning
4 min read

Green-VLA: One AI Brain for All Robots

Green-VLA trains one policy for humanoids, mobile manipulators and fixed arms across five stages and 3,000 hours of demos, hitting state of the art on CALVIN ABC-D.

AIRoboticsVLA
5 min read

To Grok Grokking: Why Neural Networks Sometimes Understand Late

Grokking happens even in ridge regression: fast memorization, a long plateau of orthogonal noise, then late generalization with a delay that scales as 1/λ.

Machine LearningGeneralizationGrokking
5 min read

Tensor Networks: A Mathematical Bridge Between Neural and Symbolic AI

Tensor networks give neural and symbolic AI one language: logic becomes sparse tensors, neural nets low-rank ones, and inference in both is a contraction.

Machine LearningNeuro-Symbolic AITensor Networks
5 min read

M²FMoE: When Experts Learn to Predict Floods

M2FMoE routes Fourier and wavelet experts by frequency band to forecast floods without extreme-event labels, cutting error 45.7% on Almaden and beating 13 baselines.

Machine LearningTime SeriesMixture of Experts
4 min read

BALLAST: When a Bandit Teaches Your Database How Long to Wait

BALLAST replaces Raft's random election timeouts with a LinUCB contextual bandit; a Rust simulator cuts cluster unavailability by 84%, and 90.8% under bursty WAN.

Machine LearningDistributed SystemsRaft
3 min read

AI Co-Scientist: Teaching Models to Write Research Plans Better Than Humans

Rubrics mined from papers turn a frozen grader into an RL reward; experts preferred the generated research plans 70% over baseline and 52% over published ones.

Machine LearningLLMReinforcement Learning
3 min read

HyDRA: Teaching Your Phone to Understand Images Without Breaking the Bank

HyDRA allocates LoRA rank per layer and inside layers automatically, buying a 4.7% gain on mobile vision-language models with no extra trainable parameters.

Machine LearningLoRAFine-tuning
3 min read

Comp-LLM: When an Army of Experts Beats a Giant – An Analysis of a Revolution in AI Architecture

Comp-LLM routes a query as a DAG of sub-queries to parallel experts: about 35B scores 0.83 F1 to Llama-2 70B's 0.85, and 1.1-1.7x faster than sequential agents.

AILLMMachine Learning
13 min read

NVIDIA Nemotron Parse v1.1: The Complete Anatomy of the Digital Document Understanding Revolution

Nemotron Parse v1.1 pairs a C-RADIO ViT-H encoder with an mBART decoder in 885M parameters, squeezing 13,184 visual tokens to 3,201 and emitting Markdown or LaTeX.

3 min read

Cost-Constrained LLM Cascades — Meet C3PO

C3PO runs LLM cascades under a probabilistic cost constraint: stay in budget with 95% confidence, escalating to a bigger model only when the cheap one falls short.

AILLMMachine Learning
3 min read

Accurate Satellite Rain Forecasting with Physics-Conditioned Neural Networks

TUPANN nowcasts rainfall from satellite data alone, adding a differentiable advection operator and lead-time-conditioned MaxViT for forecasts 10-180 minutes ahead.

AIweathermachine learning
4 min read

A Universal Crime Predictor – How Hypernetworks and Knowledge Graphs Are Transforming Forecasting

HYSTL pairs a crime knowledge graph with a hypernetwork that generates GNN weights per crime type, letting one model forecast across cities with disjoint taxonomies.

4 min read

CHORD — Smart On-Device Recommendations Without Killing Your Battery

CHORD sends each phone a bit map from a cloud hypernetwork, quantizing a sequential recommender channel by channel so weak devices keep accuracy and battery.

AIrecommendationedge computing
3 min read

DIALTREE‑RPO — a simple guide to 'Tree‑based Dialogue Reinforced Policy Optimization'

Friendly explanation of 'Tree‑based Dialogue Reinforced Policy Optimization for Red‑Teaming Attacks' — with examples and LaTeX formulas.

AIred-teamingLLM
3 min read

Attention as a Compass – Teaching Reasoning Models to Explore Smarter

AttnRL uses a model's own attention to pick branching points in process-supervised RL, beating GRPO and TreeRL by 1.8 points on math while cutting training time 8%.

AIreinforcement learningLLM
2 min read

How to Detect Credit Card Fraud?

Credit card fraud is 492 cases in 284,807 transactions; pairing undersampling with SMOTE beats either alone, and KNN and MLP gain most on the imbalanced test set.

AIfraud detectionmachine learning
3 min read

Quantum Trading – AI and Quantum Computing in Investing

A QLSTM forecaster plus a QA3C agent trade USD/TWD with just 244 parameters, returning 11.87% over 2020-2025 at 0.92% max drawdown versus classical A3C's 11.42%.

AIQuantum ComputingReinforcement Learning
2 min read

Reinforcement Learning in Pinterest Ads – DRL-PUT in action!

DRL-PUT replaces Pinterest's hand-tuned ad ranking weights with a policy network picking utility weights per request; A/B tests show CTR up 9.7%, LC-CTR up 7.7%.

reinforcement learningrecommender systemsPinterest
2 min read

Uczenie ze Wzmocnieniem w Reklamach na Pinterest – jak DRL-PUT zmienia grę / Reinforcement Learning in Pinterest Ads – How DRL-PUT Changes the Game

Pinterest's DRL-PUT learns the weights of its ad ranking utility with policy gradients instead of tuning them by hand, lifting CTR by 9.7% and LC-CTR by 7.7%.

reinforcement learningrekomendacjePinterest
4 min read

The Anatomy of AI Lies: How Language Models Can Deceive Us

Can LLMs Lie? traces deception to layers 10-15 with logit lens, zero-ablation and steering vectors: models rehearse lies in dummy tokens; bigger models lie better.

AIArtificial IntelligenceLLM
4 min read

Edge AI: How to Accelerate Neural Networks on Specialized Hardware

SLAC's SNL and Auto-SNL push neural network inference onto MPSoC FPGAs for LCLS-II, swapping weights without reconfiguration and matching or beating hls4ml latency.

AIFPGAHardware Acceleration
4 min read

Global Guarantees of Robustness: A Probabilistic Approach to AI Safety

Rather than certify every point, Mu and Lim estimate the probability that a random input is non-robust, wrapping a Clopper-Pearson interval around a small sample.

Artificial IntelligenceMachine LearningAdversarial Robustness
2 min read

Intern-S1: The New AI Scientist That's Redefining Research

Intern-S1 is an open-source MoE science foundation model with 28B activated parameters, trained on 5T tokens and a Mixture-of-Rewards spanning over 1000 tasks.

AIMachine LearningScience
6 min read

Look Inside Seamless Flow's Hyper-Efficient Training

Seamless Flow kills RL pipeline bubbles with a data plane and tag-driven GPU scheduling: up to 100% more throughput than VERL, and 23.0% to 45.8% on SWE-Bench.

AIReinforcement LearningLLM
4 min read

Systematization of Knowledge: Data Minimization in Machine Learning

SoK: Data Minimization in Machine Learning maps the GDPR principle onto ML pipelines, separating horizontal from vertical techniques and pre-hoc from post-hoc steps.

machine learningdata minimizationAI
3 min read

Learning Machines That Don't Forget: A New Method for Evolving Data

DyMoE counters catastrophic forgetting in incremental graph learning by adding a new expert per increment, with gating and a loss that makes older experts teach it.

Artificial IntelligenceMachine LearningGraphs
4 min read

Dynamic Fine-Tuning (DFT): How a Single Line of Code is Revolutionizing AI Training

Dynamic Fine-Tuning rescales each token loss by the inverse of its probability, a one-line change that beats SFT on generalization and rivals offline RL methods.

AILLMMachine Learning
5 min read

A Deep Dive into the Text-to-SQL Revolution: Analyzing the Adaptive Method

Adaptive Text-to-SQL splits the work in two: a Selector retrieves conceptually similar question-schema-SQL triples, a Generator writes the query in-context.

Artificial IntelligenceNatural Language ProcessingSQL
3 min read

ASkDAgger: How Artificial Intelligence Learns More Effectively by Asking Questions

ASkDAgger lets a robot flag an uncertain plan before acting; adaptive gating and foresight replay cut teacher queries and raise success on CLIPort and a real arm.

AIMachine LearningRobotics
3 min read

CaPulse: Teaching Machines to Hear the Rhythm of Data

CaPulse spots time-series anomalies by learning the causal rhythm of data: it fires when the cause-effect order breaks, not when a threshold is crossed.

AIMachine LearningScience
4 min read

Goedel-Prover-V2: A Revolution in Automated Theorem Proving

Goedel-Prover-V2 learns from scaffolded proof sketches and its own failed Lean attempts, outperforming AlphaProof and DeepSeek-Prover on miniF2F and PutnamBench.

AIArtificial IntelligenceMathematics
2 min read

How to Teach AI to Handle Mistakes? Meet ε-Softmax

ε-Softmax swaps rigid one-hot targets for slightly softened ones, so a mislabeled example costs the model far less. A NeurIPS 2024 answer to noisy training labels.

AIMachine LearningResearch
2 min read

Simple and Effective Method for Uncertainty Quantification

Uncertainty from one deterministic model: kernel density estimation in feature space builds an information potential field flagging OOD samples on CIFAR-10 vs SVHN.

Machine LearningUncertainty QuantificationOOD Detection
2 min read

Deep Learning-based Prediction of Clinical Trial Enrollment with Uncertainty Estimates

A multimodal Clinical Longformer model predicts trial enrollment at R² 0.76 and forecasts recruitment duration with 7.52-month MAE versus 10.55 for fit-and-filter.

clinical trialsmachine learningdeep learning
1 min read

Consensus-Driven Active Model Selection

CODA picks the best model from many candidates via their agreement: Bayesian Dawid-Skene plus information gain wins 18 of 26 benchmarks with up to 70% fewer labels.

machine learningactive learningmodel selection
1 min read

RLVMR: Reinforcement Learning with Verifiable Meta‑Reasoning Rewards for Robust Long‑Horizon Agents

RLVMR adds a verifiable meta-reasoning reward to the environment reward, scoring the agent's reasoning trace so long-horizon agents take fewer semantic shortcuts.

AIadaptacyjna sensorykarle
2 min read

How AI Can Reveal Where Your Honey Comes From — A Look at Mineral Fingerprints

Random Forests read the 12-mineral fingerprint of 429 honey samples to name the source flower with 99.50% accuracy and the region of origin with 98.01%.

honeymachine-learningfood-authenticity
1 min read

Optimizing Call Center Operations with Reinforcement Learning: PPO vs. Value Iteration

PPO beats classical Value Iteration at call routing in a discrete-event call centre simulation: shortest customer wait, least agent idle time, highest RL reward.

Reinforcement LearningPPOValue Iteration
2 min read

Efficient & Geometrically-Smart: Linear Memory SE(2)-Invariant Attention Explained

Fourier-encoded relative pose lets SE(2)-invariant attention use linear rather than quadratic memory, with error below 0.001. Best paper at the RSS 2025 workshop.

Machine LearningTransformersGeometric AI
2 min read

A Lightweight AI Engine for Skin Cancer Detection on Wearable Devices

A MobileNetV2 skin-lesion classifier compressed with TensorRT runs on a Jetson Orin Nano at 41% of its original size, scoring 87.2% F1 with 93% lower power in INT8.

AIMobileNetV2skin cancer
2 min read

SOPHIA: Enhancing Slow‑Thinking in Large Vision‑Language Models

SOPHIA mixes on-policy visual grounding with off-policy reasoning chains and backward reward propagation, adding 8.5 points of average accuracy to InternVL3.0-38B.

AILVLMreinforcement learning
2 min read

The Role of AI in Managing Satellite Constellations

ConstellAI, an ESA project, uses reinforcement learning for routing in satellite mega-constellations, beating Dijkstra on delay and predicting bandwidth demand.

ConstellAIReinforcement LearningSatellites
1 min read

On the Fundamental Limitations of Dual Static CVaR Decompositions in Markov Decision Processes

Godbout and Durand show the two dual CVaR formulas can share no feasible risk assignment, and build an MDP where no policy is optimal at every risk level.

Markov Decision ProcessesCVaRdual decomposition
1 min read

PinFM: Foundation Model for User Activity Sequences at a Billion-Scale Visual Discovery Platform

PinFM, Pinterest's 20B-parameter sequence transformer, uses deduplicated cross-attention and int4 embeddings to lift throughput 600% and engagement 20%.

2 min read

GradNetOT: Learning Optimal Transport Maps with GradNets

GradNetOT learns optimal transport maps as the gradient of a convex potential, using a monotone gradient network trained on the Monge–Ampère residual.

GradNetOTOptimal TransportNeural Networks
1 min read

Unstable Power: How Sharpness Drives Deep Network Learning

A larger learning rate leaves the NTK better aligned with the labels at the Edge of Stability: Kernel Target Alignment jumps exactly when sharpness drops.

popularizationmachine learningNTK
2 min read

RiemannLoRA: A Unified Riemannian Framework for Ambiguity-Free LoRA Optimization

RiemannLoRA places LoRA updates on the manifold of rank-r matrices, erasing factorization ambiguity: half the fine-tuning steps and up to 1.2 BLEU more.

Machine LearningRiemannian OptimizationLoRA
1 min read

A Neural Network Model of Complementary Learning Systems: Pattern Separation and Completion for Continual Learning

A complementary learning systems model pairs a VAE with a modern Hopfield network for pattern separation and completion, hitting 90% average accuracy on Split-MNIST.

1 min read

Target Polish: How to Polish Data and Reveal Its True Structure

Target Polish cleans data before factorizing it: clip X into a delta band around the current WH estimate, refactorize, repeat. Robust NMF for matrices and tensors.

NMFartificial intelligencemachine learning
1 min read

Optimistic Exploration for Risk-Averse Constrained Reinforcement Learning

ORAC maximizes an upper confidence bound on reward while minimizing a lower bound on cost, with an adaptive penalty, tested on Safety-Gymnasium and CityLearn.

1 min read

Not Just Bigger Models: Why AI Should See Better Instead of Just Scaling

Adaptive sensing beats brute-force scale: EfficientNet-B0 paired with sensors that tune themselves to the scene outperforms OpenCLIP-H on perception tasks.

AIadaptive sensingmachine learning
3 min read

HGMP: Revolutionizing Complex Graph Analysis with Prompt Learning

HGMP recasts node, edge and ranking tasks as graph-level prompts, combining contrastive pre-training with per-type feature prompts for heterogeneous graph learning.

graph learningprompt learningmulti-task
2 min read

Predicting and Generating Antibiotics Against Future Pathogens with ApexOracle

ApexOracle fuses Evo2 genome embeddings with a SELFIES diffusion model to predict MIC and design new antibiotics, with R2 gains up to 15% on P. aeruginosa.

MLAntibioticsDrug Discovery
2 min read

HeLo – A New Path for Multimodal Emotion Recognition

HeLo predicts a distribution over emotions instead of one label, fusing ECG and video via cross-attention, entropic optimal transport and a label-correlation loss.

multimodalemotionmachine learning
1 min read

Modern Methods in Associative Memory

Hopfield networks store only 0.14N patterns; dense variants reach polynomial capacity and modern ones exponential, with a Lagrangian view linking them to attention.

Associative MemoryHopfield NetworksTransformers
2 min read

QuEst: Blending Data and Predictions for Robust Quantile Estimation

QuEst blends a small gold sample with large model-generated imputations to estimate quantiles and CVaR, canceling simulator bias and minimizing sampling variance.

QuEstQuantile EstimationStatistics
2 min read

RetrySQL: Self-Correcting Query Generation

RetrySQL injects deliberate wrong reasoning steps and a BACK token into text-to-SQL pretraining, lifting OpenCoder 1.5B execution accuracy by 4 points toward GPT-4o.

RetrySQLtext-to-SQLLLM
2 min read

How Modern Information Theory Helps Diagnose Mental Disorders – MvHo‑IB in Action

MvHo-IB diagnoses brain disorders from rs-fMRI by adding triplet O-information to pairwise connectivity, gaining up to 7% accuracy on ABIDE, ADHD-200 and COBRE.

fMRIAIdeep learning
2 min read

Multi-level Stepwise Hints in Reinforcement Learning

StepHint splits reasoning chains by end-of-step token probability, then feeds RL agents hints at several levels, easing near-miss errors and stalled exploration.

1 min read

How to Predict Scooter Demand? XGBoost and Urban Micromobility

XGBoost on 341 temporal, spatial and network features predicts e-scooter demand 27-49% better than rivals and up to 90% better than Prophet, node centrality on top.

XGBoostAImachine learning
1 min read

Ghost Nodes: A Trick That Makes Neural Networks Learn Smarter

Ghost nodes add fake output classes: softmax runs over 13 outputs while loss counts only the 10 real ones, giving gradients room to escape flat regions early.

ghostAImachine learning
2 min read

Does artificial intelligence really understand math? Let's find out what it says... data audit?

A neural-mechanistic hybrid forecasts epidemics from differentially private mobility and spending data, still beating public-health-only models under strong noise.

prywatnośćAInauczanie maszynowe
1 min read

Unbreakable in the Face of Adversity: ARMOR – Resilient UAV Control

ARMOR trains a teacher VAE on privileged attack data and an LSTM student on raw sensor history, giving UAVs zero-shot robustness to unseen GPS spoofing and jamming.

2 min read

Mind2Web 2: A new era of “agent-based” web search

Mind2Web 2 grades agentic web search on 130 tasks with an Agent-as-a-Judge rubric; the best of nine systems reaches 50-70% of human quality at twice the speed.

AIsearchingmind2web
2 min read

A Machine That Discovers the Laws of Physics: How H-FEX Works and Why It Matters

H-FEX searches symbolic expressions for a system's Hamiltonian, recovering exp(-1.0003p² - 1.1002q⁴) almost exactly and beating SINDy on long-term energy drift.

physicsAIsymbolic learning
3 min read

When the Bandit Is Stronger Than Your Model – On the Limits of Exploratory Learning

Bandit learning has no VC-dimension analogue: a class can need only 2 queries yet stay intractable unless RP = NP. With a Rust epsilon-greedy A/B/n example.