Graph Crimes of the Temporal Kind: How LoReTTA Quietly Breaks Time
LoReTTA shows that temporal graph models can be weakened by subtle, resource-light poisoning of interaction history, not only by noisy brute-force attack.
Cognaptus Insights
Industry thoughts, trends, and commentary.
Recurrent Revival: How Retrofitted Depth Turns LLMs Into Deeper Thinkers
A mechanism-first look at how pretrained language models can be surgically converted into depth-recurrent reasoners—and why the gains are useful, conditional, and not remotely free.
Replan, Rethink, Repeat: Why Vision-Language Models Make Better Closed‑Loop Planners
A comparison-led reading of why VLM robot planners need feedback, memory, and carefully tuned replanning rather than blind faith in more model calls.
Scalpels, Agents, and Orchestrators: When Surgery Meets Autonomous Workflows
A mechanism-first reading of how VISA uses LLM orchestration, memory, and specialised agents to make voice-controlled surgical data interfaces more reliable.
Think Outside the Bounding Box: How SpatialThinker Reinforces 3D Reasoning
SpatialThinker shows that better reward design, not just more data or depth sensors, can make multimodal models reason more reliably about 3D space.
When Noisy Data Talks Back: The Fragile Art of Learning Under Infinite Contamination
A practical reading of new learning-theory results showing why noisy data may still permit generation, but can quietly destroy broad coverage.
When Videos Grow Hands: How PhysWorld Teaches Robots to Stop Hallucinating Physics
PhysWorld shows how generated task videos become useful for robots only after geometry, physics, and residual learning do the unfashionable work.
Back to the Drawing Board: How DiagramIR Quietly Fixes Math Diagrams for AI
A mechanism-first look at why DiagramIR’s structured back-translation pipeline makes AI-generated math diagrams easier to verify, cheaper to govern, and harder to excuse.
Charts Without Tears: When AI Starts Cleaning Your Data So You Don’t Have To
A close reading of an AI data-visualization platform paper shows where automated analytics can compress workflow, and where the evidence still stops short of replacing analysts.
GraphRAG Gone Modular: Why Multi-Agent Cypher Matters More Than You Think
A mechanism-first analysis of how multi-agent Text-to-Cypher systems turn graph querying from brittle prompting into executable, database-grounded retrieval.