AI × Crypto News Feed

Chinese Cyberbullying Detection: Dataset, Method, and Validation

Expert Evaluation and the Limits of Human Feedback in Mental Health AI Safety Testing

Simultaneous Long-tailed Recognition and Multi-modal Fusion for Highly Imbalanced Multi-modal Data

Can We Formally Verify Neural PDE Surrogates? SMT Compilation of Small Fourier Neural Operators

Researchers demonstrate that Fourier Neural Operators (FNOs) used for PDE simulation can be formally verified using SMT solvers by exploiting their piecewise-linear structure once weights are fixed. While exact encoding provides sound proofs and counterexamples on small models, scalability remains limited, revealing a fundamental tradeoff between formal verification rigor and practical applicability for production neural operators.

CDS4RAG: Cyclic Dual-Sequential Hyperparameter Optimization for RAG

Researchers introduce CDS4RAG, a novel optimization framework that improves Retrieval-Augmented Generation systems by cyclically optimizing retriever and generator hyperparameters separately rather than treating them as a monolithic unit. The method achieves up to 1.54x improvements in generation quality while demonstrating faster convergence across multiple benchmarks and language models.

NeuroGAN-3D: Enhancing Intrinsic Functional Brain Networks via High-Fidelity 3D Generative Super-Resolution

Researchers have developed NeuroGAN-3D, a generative AI model that enhances the spatial resolution of functional brain imaging maps derived from resting-state fMRI scans. The technology leverages adversarial neural networks to improve the precision of neuroimaging data, enabling better detection of brain connectivity patterns and potential biomarkers for neurological conditions.

Bridging Sequence and Graph Structure for Epigenetic Age Prediction

Researchers present a novel machine learning framework that combines DNA sequence analysis with graph neural networks to predict biological age from methylation patterns, achieving 12.8% improvement over existing methods. The approach uses handcrafted sequence features rather than deep learning to encode biological context, demonstrating practical advantages in aging research applications.

ProteinOPD: Towards Effective and Efficient Preference Alignment for Protein Design

DynGhost: Temporally-Modelled Transformer for Dynamic Ghost Imaging with Quantum Detectors

Interactive Critique-Revision Training for Reliable Structured LLM Generation

Researchers propose DPA-GRPO, a novel training method for large language models that improves structured decision-making by using a generator-verifier framework where one model produces outputs and another validates them through safety assurance cases. The method demonstrates improved accuracy on tax calculation benchmarks and addresses the challenge of ensuring LLM outputs are locally correct, globally consistent, and auditable.

Optimal FALQON for Quantum Approximate Optimization via Layer-wise Parameter Tuning

Researchers present Optimal FALQON, an enhanced quantum optimization algorithm that adaptively tunes layer-wise parameters to improve performance on noisy quantum devices. Testing on 3-regular graphs demonstrates significant improvements in convergence speed and solution quality compared to standard approaches, with implications for practical quantum computing applications.

LLM Advertisement based on Neuron Auctions

Researchers introduce Neuron Auctions, a novel mechanism that embeds advertisements within Large Language Models by targeting their internal neural representations rather than surface text. The approach uses mechanistic interpretability to identify brand-specific neurons that operate in near-orthogonal subspaces, enabling platforms to balance advertiser revenue, user experience, and content quality through a strategy-proof auction mechanism.

CAMAL: Improving Attention Alignment and Faithfulness with Segmentation Masks

Researchers introduce CAMAL, a method that leverages segmentation masks to improve attention alignment and faithfulness in vision models across deep learning and reinforcement learning paradigms. The approach achieves over 35% improvements in attention faithfulness while maintaining or improving generalization performance without additional inference costs.

SearchSkill: Teaching LLMs to Use Search Tools with Evolving Skill Banks

SearchSkill is a new framework that teaches language models to perform more effective web searches by explicitly planning queries through reusable skill cards rather than treating search as an undifferentiated action. The system maintains an evolving skill bank that improves from failure patterns, demonstrating better performance on knowledge-intensive QA tasks with fewer wasted queries and improved reasoning accuracy.

CADBench: A Multimodal Benchmark for AI-Assisted CAD Program Generation

A Cold Diffusion Approach for Percussive Dereverberation

Empty SPACE: Cross-Attention Sparsity for Concept Erasure in Diffusion Models

Reinforcement Learning for Scalable and Trustworthy Intelligent Systems

A dissertation presents research on scaling reinforcement learning across distributed systems while ensuring trustworthy behavior in AI applications. The work addresses communication efficiency in federated settings and alignment with human preferences in large language models, proposing that next-generation intelligent systems require both optimization efficiency and safety mechanisms.

Internalizing Safety Understanding in Large Reasoning Models via Verification

Researchers propose Safety Internal (SInternal), a framework that trains large reasoning models to verify the safety of their own outputs rather than relying on external compliance mechanisms. The approach demonstrates that models can internalize safety understanding through verification tasks, significantly improving robustness against adversarial jailbreaks and out-of-domain attacks.

SDG-MoE: Signed Debate Graph Mixture-of-Experts

Researchers introduce SDG-MoE, a novel mixture-of-experts architecture that enables deliberation among routed experts through signed graph communication before output aggregation. The model demonstrates 19.8% perplexity improvement over vanilla MoE and achieves state-of-the-art results on multiple language modeling benchmarks while maintaining computational efficiency.

FQPDR: Federated Quantum Neural Network for Privacy-preserving Early Detection of Diabetic Retinopathy

Researchers propose FQPDR, a federated quantum neural network system for early detection of diabetic retinopathy that preserves patient privacy by processing medical data locally rather than centralizing it. The approach combines federated learning with quantum computing to identify microaneurysm dots—the earliest signs of diabetic retinopathy—while maintaining data confidentiality across distributed healthcare systems.

Ace-Skill: Bootstrapping Multimodal Agents with Prioritized and Clustered Evolution

Researchers introduce Ace-Skill, a co-evolutionary framework that improves multimodal AI agents by optimizing both data sampling and knowledge organization. The system achieves 35% performance gains on tool-use benchmarks and enables smaller models to inherit capabilities from larger ones without additional training.

M$^3$: Reframing Training Measures for Discretized Physical Simulations

Researchers introduce M³ (Multi-scale Morton Measure), a framework that improves neural surrogate models for physical simulations by addressing training bias from discretized data sampling. The method achieves up to 4.7× error reduction in volumetric cases and maintains superior performance even with 90% data reduction, demonstrating that data distribution strategy significantly impacts operator learning efficiency.

Mazocarta: A Seeded Procedural Deckbuilder for Instrumented Game Development

Mazocarta is an open-source procedural deckbuilder game built in Rust and WebAssembly that serves as a reference architecture for instrumented game development. The project demonstrates how a single rules engine can support interactive play, automated testing, balance simulation, and local multiplayer, with evaluation showing 36.1% single-player and 34.9% two-player win rates across 1,000 deterministic seeds.

HeteroGenManip: Generalizable Manipulation For Heterogeneous Object Interactions