Google Research Meets Bayesian Deep Learning: Shared States, Kernels, and PFNs
Neural Trend Hub
Google Research Meets Bayesian Deep Learning: Shared States, Kernels, and PFNs
6:05
Memory-Efficient LLMs: Attention I/O, KV Cache Eviction, and MoE Compression
Neural Trend Hub
Memory-Efficient LLMs: Attention I/O, KV Cache Eviction, and MoE Compression
6:28
Diffusion and Dynamics: From Finite-Particle Learning to Survival Prediction
Neural Trend Hub
Diffusion and Dynamics: From Finite-Particle Learning to Survival Prediction
6:39
Qwen, Yale, and the New Retrieval Frontier: Syntax, Mood, and In-Context Memory
Neural Trend Hub
Qwen, Yale, and the New Retrieval Frontier: Syntax, Mood, and In-Context Memory
6:12
Breaking Bottlenecks in Multimodal Training, Multilingual Reasoning, and Self-Improvement
Neural Trend Hub
Breaking Bottlenecks in Multimodal Training, Multilingual Reasoning, and Self-Improvement
6:33
Discrete Diffusion Alignment: Reward-Tilted Sampling, Gibbs Correctors, and Few-Step Control
Neural Trend Hub
Discrete Diffusion Alignment: Reward-Tilted Sampling, Gibbs Correctors, and Few-Step Control
6:14
Beyond Invariance: How Contrastive Structure Shapes Learning in Images, Time Series, and Depth
Neural Trend Hub
Beyond Invariance: How Contrastive Structure Shapes Learning in Images, Time Series, and Depth
6:08
Spectral Optimizers and Stability: Robust Learning Theory for Modern Training
Neural Trend Hub
Spectral Optimizers and Stability: Robust Learning Theory for Modern Training
6:28
Learning and Control in the Age of Overparameterization and Multi-Agent Interaction
Neural Trend Hub
Learning and Control in the Age of Overparameterization and Multi-Agent Interaction
5:51
From Hidden Circuits to Uncertain Predictions: Reliability in Modern ML Models
Neural Trend Hub
From Hidden Circuits to Uncertain Predictions: Reliability in Modern ML Models
6:44
Beyond Static Data: Self-Play, Simulations, and Hysteretic Memory in AI
Neural Trend Hub
Beyond Static Data: Self-Play, Simulations, and Hysteretic Memory in AI
6:13
From Lung Pathology to Peer Review: AI Models, Limits, and Scientific Insight
Neural Trend Hub
From Lung Pathology to Peer Review: AI Models, Limits, and Scientific Insight
7:04
From Concepts to Streams: Unified Multimodal Pretraining and Editing
Neural Trend Hub
From Concepts to Streams: Unified Multimodal Pretraining and Editing
5:57
From Sparse Experts to Convex Tokenizers: Certified Structure in NLP
Neural Trend Hub
From Sparse Experts to Convex Tokenizers: Certified Structure in NLP
6:49
Memory-Savvy Inference: Portable LLMs, Private Trees, and Verified KV Caches
Neural Trend Hub
Memory-Savvy Inference: Portable LLMs, Private Trees, and Verified KV Caches
6:10
When LLMs Learn, Simulate, and Diffuse: Hidden Biases in Generative Models
Neural Trend Hub
When LLMs Learn, Simulate, and Diffuse: Hidden Biases in Generative Models
6:42
Semantic Diffusion Across Video, Images, and Audio: Faster, Smarter Generation
Neural Trend Hub
Semantic Diffusion Across Video, Images, and Audio: Faster, Smarter Generation
6:01
Beyond One-Size-Fits-All: Bias, Layerwise Learning Rates, and SGD Dynamics
Neural Trend Hub
Beyond One-Size-Fits-All: Bias, Layerwise Learning Rates, and SGD Dynamics
6:13
Transport, Flow, and Memorization: Geometry and Generalization in Generative Models
Neural Trend Hub
Transport, Flow, and Memorization: Geometry and Generalization in Generative Models
6:32
Unifying Robust Representation Learning Across Time Series, Domain Shift, and Wearables
Neural Trend Hub
Unifying Robust Representation Learning Across Time Series, Domain Shift, and Wearables
7:01
Limits, Dynamics, and Hidden Computation in Transformers
Neural Trend Hub
Limits, Dynamics, and Hidden Computation in Transformers
6:46
Efficient, Universal, and Better-Conditioned Neural Architectures
Neural Trend Hub
Efficient, Universal, and Better-Conditioned Neural Architectures
6:24
Memory, Mind, and Resilience in Modern LLM Systems
Neural Trend Hub
Memory, Mind, and Resilience in Modern LLM Systems
6:32
Unifying Audits, Memory, and Neighborhoods for Safer Distributed AI
Neural Trend Hub
Unifying Audits, Memory, and Neighborhoods for Safer Distributed AI
6:08
Rubrics, Verification, and the Mirage of Reliable Self-Critique in AI Systems
Neural Trend Hub
Rubrics, Verification, and the Mirage of Reliable Self-Critique in AI Systems
6:34
Beyond Scaling and Similarity: Collaboration, Inference, and Test-Driven AI
Neural Trend Hub
Beyond Scaling and Similarity: Collaboration, Inference, and Test-Driven AI
6:23
Compressing AI at the Edge: Distributed Inference, 2-Bit Caches, and Tiny Students
Neural Trend Hub
Compressing AI at the Edge: Distributed Inference, 2-Bit Caches, and Tiny Students
6:43
From Visual Thought to Dorsal Control: Multimodal Models That See, Act, and Measure
Neural Trend Hub
From Visual Thought to Dorsal Control: Multimodal Models That See, Act, and Measure
6:39
Beyond Visual Polish: Benchmarking Reasoning in World Models and Coding Agents
Neural Trend Hub
Beyond Visual Polish: Benchmarking Reasoning in World Models and Coding Agents
7:05
Agentic Memory and Attribution in Multimodal AI and Secure Microservices
Neural Trend Hub
Agentic Memory and Attribution in Multimodal AI and Secure Microservices
6:29
On-Policy Learning for Agents: Flow, Search, and Tool Use in the Real World
Neural Trend Hub
On-Policy Learning for Agents: Flow, Search, and Tool Use in the Real World
6:32
Agent Reliability Under Adversarial Context, Real Tools, and Generative Dynamics
Neural Trend Hub
Agent Reliability Under Adversarial Context, Real Tools, and Generative Dynamics
6:38
Alignment, Infrastructure, and World Models: Building Agents That Persist
Neural Trend Hub
Alignment, Infrastructure, and World Models: Building Agents That Persist
7:11
Beyond Robustness: Agentic Optimization, Semantic Attacks, and Quantization Backdoors
Neural Trend Hub
Beyond Robustness: Agentic Optimization, Semantic Attacks, and Quantization Backdoors
6:41
Hidden Paths in AI: Context Limits, Scraper Attribution, and Finetuning Evasion
Neural Trend Hub
Hidden Paths in AI: Context Limits, Scraper Attribution, and Finetuning Evasion
6:02
Benchmarking Intelligence Across Voice Agents, Quantum Circuits, and Multimodal Memory
Neural Trend Hub
Benchmarking Intelligence Across Voice Agents, Quantum Circuits, and Multimodal Memory
6:07
Beyond Text: Benchmarking Real-World Failure Modes in AI Agents and Medical Synthesis
Neural Trend Hub
Beyond Text: Benchmarking Real-World Failure Modes in AI Agents and Medical Synthesis
6:28
When Models Learn the Wrong Lesson: Negation, Personalization, and Rollout Truth
Neural Trend Hub
When Models Learn the Wrong Lesson: Negation, Personalization, and Rollout Truth
6:54
From Solver-Grounded Multimodal Models to Robust and Efficient Learning
Neural Trend Hub
From Solver-Grounded Multimodal Models to Robust and Efficient Learning
6:23
From Brainwaves to Bias: Multimodal Models, Hidden Harms, and Alignment Under Pressure
Neural Trend Hub
From Brainwaves to Bias: Multimodal Models, Hidden Harms, and Alignment Under Pressure
7:18
Bridging Distribution Gaps: Diffusion Distillation, Simulators, and Expert Routing
Neural Trend Hub
Bridging Distribution Gaps: Diffusion Distillation, Simulators, and Expert Routing
7:08
Benchmarks for Trustworthy AI: Evidence, Grounding, and Scientific Judgment
Neural Trend Hub
Benchmarks for Trustworthy AI: Evidence, Grounding, and Scientific Judgment
7:03
Beyond Surface Patterns: Exact Constraints, Dialect Shifts, and Symbolic Semantics
Neural Trend Hub
Beyond Surface Patterns: Exact Constraints, Dialect Shifts, and Symbolic Semantics
6:48
When AI Learns, Diagnoses, and Flatters: Alignment, Resilience, and Relational Costs
Neural Trend Hub
When AI Learns, Diagnoses, and Flatters: Alignment, Resilience, and Relational Costs
6:33
Verifiable AI: Trust, Retrieval, and Physics-Preserving Structure in Modern Models
Neural Trend Hub
Verifiable AI: Trust, Retrieval, and Physics-Preserving Structure in Modern Models
6:37
From Instrumental Agents to Self-Normalized Learning: Unified Frontiers in AI Behavior
Neural Trend Hub
From Instrumental Agents to Self-Normalized Learning: Unified Frontiers in AI Behavior
6:54
From Evidence Gates to Stateful Memory: Reliable AI Across Discovery and Generation
Neural Trend Hub
From Evidence Gates to Stateful Memory: Reliable AI Across Discovery and Generation
6:32
Safety, Modularity, and Robustness in Modern AI Systems
Neural Trend Hub
Safety, Modularity, and Robustness in Modern AI Systems
6:41
When Simplicity Wins: Cost-Efficient Serving, Invariant Diagnostics, and Network Inference
Neural Trend Hub
When Simplicity Wins: Cost-Efficient Serving, Invariant Diagnostics, and Network Inference
6:56
Beyond Bigger Models: Benchmarks, Sparse Evidence, and Control in Modern AI
Neural Trend Hub
Beyond Bigger Models: Benchmarks, Sparse Evidence, and Control in Modern AI
6:41
Beyond Raw Scale: Agentic Interpretability, Quantum Learning, and Efficient Serving
Neural Trend Hub
Beyond Raw Scale: Agentic Interpretability, Quantum Learning, and Efficient Serving
6:53
From AI Red Teaming to Optimal Allocation and Regret: Unified Decision-Making
Neural Trend Hub
From AI Red Teaming to Optimal Allocation and Regret: Unified Decision-Making
6:28
Beyond Dense Data: Foundation Models, Sparse Views, and Residual Evidence
Neural Trend Hub
Beyond Dense Data: Foundation Models, Sparse Views, and Residual Evidence
7:04
Benchmarks That Expose Hidden Failure Modes in AI Systems
Neural Trend Hub
Benchmarks That Expose Hidden Failure Modes in AI Systems
6:54
Beyond the Prompt: Benchmarks for Multihop Reasoning, Support, and Discovery
Neural Trend Hub
Beyond the Prompt: Benchmarks for Multihop Reasoning, Support, and Discovery
7:04
Structured Intelligence in Action: Graphs, Operators, and Onchain Agents
Neural Trend Hub
Structured Intelligence in Action: Graphs, Operators, and Onchain Agents
6:53
Beyond Fidelity and Prompts: Observable Evolution, Downstream Utility, and Fairness in AI Systems
Neural Trend Hub
Beyond Fidelity and Prompts: Observable Evolution, Downstream Utility, and Fairness in AI Systems
8:15
From Search Drift to Synchronization and CT-Constrained Motion: Structured Intelligence
Neural Trend Hub
From Search Drift to Synchronization and CT-Constrained Motion: Structured Intelligence
8:05
Reliable State, Real-World Evaluation, and Better Generative Metrics
Neural Trend Hub
Reliable State, Real-World Evaluation, and Better Generative Metrics
7:32
Beyond Surface Signals: Evaluation, Generative Modeling, and Symmetry-Aware Diffusion
Neural Trend Hub
Beyond Surface Signals: Evaluation, Generative Modeling, and Symmetry-Aware Diffusion
8:13
Beyond Surface Signals: Evaluation, Generative Modeling, and Symmetry-Aware Diffusion
Neural Trend Hub
Beyond Surface Signals: Evaluation, Generative Modeling, and Symmetry-Aware Diffusion
7:54
Beyond One-Size-Fits-All AI: Bias, Embodiment, and Intersectional Harm
Neural Trend Hub
Beyond One-Size-Fits-All AI: Bias, Embodiment, and Intersectional Harm
8:01
Agent-Native Intelligence from Skill Retrieval to Reproducible Research Artifacts
Neural Trend Hub
Agent-Native Intelligence from Skill Retrieval to Reproducible Research Artifacts
6:00
From Geometry to Retrieval: Unifying Structured 3D Understanding Across Motion, Bodies
Neural Trend Hub
From Geometry to Retrieval: Unifying Structured 3D Understanding Across Motion, Bodies
7:52
Unifying Hidden Signals: Domain Shifts, Backdoors, and Time in Vision Models
Neural Trend Hub
Unifying Hidden Signals: Domain Shifts, Backdoors, and Time in Vision Models
7:14
From Memory to World Models to Symbols: Learning to Generalize Under Uncertainty
Neural Trend Hub
From Memory to World Models to Symbols: Learning to Generalize Under Uncertainty
7:08
Robust AI Systems from Distributed Training to Verbal and Visual Evaluation
Neural Trend Hub
Robust AI Systems from Distributed Training to Verbal and Visual Evaluation
7:45
Unifying Perspectives on Hidden Structure, Domain Adaptation, and Generalist Vision
Neural Trend Hub
Unifying Perspectives on Hidden Structure, Domain Adaptation, and Generalist Vision
7:52
Spatial Intelligence Adversarial Robustness, Geometry-Aware AI: New Frontiers in Multimodal Systems
Neural Trend Hub
Spatial Intelligence Adversarial Robustness, Geometry-Aware AI: New Frontiers in Multimodal Systems
7:34
From Geometry to Confidence: Transformers for 3D Understanding and Reliable Reasoning
Neural Trend Hub
From Geometry to Confidence: Transformers for 3D Understanding and Reliable Reasoning
7:02
Scaling Intelligence at Test Time: From Scientific Discovery to Instant Collaboration
Neural Trend Hub
Scaling Intelligence at Test Time: From Scientific Discovery to Instant Collaboration
7:42
Unifying Stable Discrete Optimization: Quantization, Diffusion, and Weakly Supervised Reasoning
Neural Trend Hub
Unifying Stable Discrete Optimization: Quantization, Diffusion, and Weakly Supervised Reasoning
7:14
Multimodal Intelligence Under the Microscope: Healthcare, Safety, and Web Coding
Neural Trend Hub
Multimodal Intelligence Under the Microscope: Healthcare, Safety, and Web Coding
8:49
Efficient Inference and Reasoning from Exascale Models to Interpretable Algorithms
Neural Trend Hub
Efficient Inference and Reasoning from Exascale Models to Interpretable Algorithms
8:20
Unifying Perspectives on Efficient Depth, Reasoning, and Representation Gaps in Modern Neural Models
Neural Trend Hub
Unifying Perspectives on Efficient Depth, Reasoning, and Representation Gaps in Modern Neural Models
7:48
Symbolic Search to Latent Planning: Structural Breakthroughs in Optimization, Robotics
Neural Trend Hub
Symbolic Search to Latent Planning: Structural Breakthroughs in Optimization, Robotics
7:21
Bridging Model Gaps: Distillation, Alignment, and Student-Consistent Reasoning
Neural Trend Hub
Bridging Model Gaps: Distillation, Alignment, and Student-Consistent Reasoning
5:19
Claude Code | World Model Efficient Optimization, Agent Design, and Hierarchical Planning
Neural Trend Hub
Claude Code | World Model Efficient Optimization, Agent Design, and Hierarchical Planning
6:57
DeepMind | Kimi | From KVCache to Consciousness: Verified Computation and Scalable AI Systems
Neural Trend Hub
DeepMind | Kimi | From KVCache to Consciousness: Verified Computation and Scalable AI Systems
8:08
Unifying World Models, Stable Recurrence, and Closed-Loop Control in Modern AI
Neural Trend Hub
Unifying World Models, Stable Recurrence, and Closed-Loop Control in Modern AI
7:38
Transferable Memory, Generative World Modeling, and Self-Evolving Spatial Intelligence
Neural Trend Hub
Transferable Memory, Generative World Modeling, and Self-Evolving Spatial Intelligence
7:42
From Benchmarks to Production: Evaluating, Diagnosing, and Scaling Agentic AI
Neural Trend Hub
From Benchmarks to Production: Evaluating, Diagnosing, and Scaling Agentic AI
8:29
Beyond Heuristics: Stable Learning From Quantization to World Models and Self-Improving LMs
Neural Trend Hub
Beyond Heuristics: Stable Learning From Quantization to World Models and Self-Improving LMs
7:30
Deep Dive: Efficient Adaptation, Self-Improving Reasoning, and Long-Context Memory in LLMs
Neural Trend Hub
Deep Dive: Efficient Adaptation, Self-Improving Reasoning, and Long-Context Memory in LLMs
8:59
From Attention to Compression: Learning the Hidden Runtime of AI
Neural Trend Hub
From Attention to Compression: Learning the Hidden Runtime of AI
7:58
From Frequency to Reasoning: How Training Shape Drives LLM Performance
Neural Trend Hub
From Frequency to Reasoning: How Training Shape Drives LLM Performance
7:50
From Live-Web Agents to Evolving Reasoning
Neural Trend Hub
From Live-Web Agents to Evolving Reasoning
8:00