FeedSources
AS

Amazon Science Publications

49 articles total

Go to source

Amazon is a great place to practice science and have real business impact with publications. Amazon scientists continue to publish papers, teach, and engage with the worldwide research community.

Go to source
  •  Scalable visual attribute recognition in e-commerce products via automated synthetic label generation
  •  The price of knowledge: Optimal algorithms for costly bandits
  •  BRIDGE: Building representations in domain-guided program synthesis
  •  Is GraphRAG needed? From basic RAG to graph-/agentic solutions with context optimization
  •  LLM-assisted diagnostics for simulation–reality discrepancies in manufacturing systems
  •  VOFA: Visual object goal pushing with force-adaptive control for humanoids
  •  Agentic translation of C to Rust
  •  Multi-phase vision-based navigation and inspection for legged robots with online goal refinement and vision-only halting
  •  DiTailed: Ensuring visual object consistency in text-image-to-image flow matching models
  •  Impact-driven event embeddings for context-aware forecasting and anomaly detection in financial time series
  •  Why is this on my bill? Documentation-grounded rule learning for itemized bill explanation
  •  Enhancing part-level point grounding for any open-source MLLMs
  •  SOP-Bench: Complex industrial SOPs for evaluating LLM agents
  •  APRS: A unified framework of multi-pricing demand simulation for Amazon device dependent products
  •  A comprehensive taxonomy of temporal dimensions in natural language queries
  •  Agentic AI for subsurface intelligence: Automated SEG-Y to MDIO migration at scale
  •  From Guess2Graph: When and how can unreliable experts safely boost causal discovery in finite samples?
  •  Self-improvement for fast, high-quality plan generation
  •  Advancing multimodal LLMs by large-scale 3D visual instruction dataset generation
  •  DACO: Dictionary-aligned concept control for safeguarding multimodal LLMs
  •  Bootstrap-conditioned action selection with tabular foundation models
  •  Library drift: Diagnosing and fixing a silent failure mode in self-evolving LLM skill libraries
  •  Time-aware prior fitted networks for zero-shot forecasting with exogenous variables
  •  Forecasting with factor-augmented time series foundation models
  •  Stress tests REVEAL fragile temporal and visual grounding in video-language models
  •  Seekable OCI: Lazy-loading container images via range-request indexing
  •  Confidence-aware multi-agent orchestration for evaluating multimodal rule compliance
  •  Composite-attribute person re-identification via pose-guided disentanglement
  •  Diagnostic knowledge graphs: Automated benchmark construction and deterministic evaluation for multi-step reasoning agents
  •  Deployment risk assessment using diff-aware features: A case study at Prime Video
  •  Super weights in LLMs and the failure of selective training
  •  Directional alignment mitigates reward hacking in reinforcement learning for language models
  •  Benchmarking deflection and hallucination in large vision-language models
  •  ProFraudGuard: Proactive adversarial fine-tuning of fraud detectors with inverse reinforcement learning
  •  A unified approach to interpretable causal root cause attribution
  •  Multi-objective reinforcement learning for large-scale tote allocation in human-robot collaborative fulfillment centers
  •  Does the optimal hallucination detector for agentic tool calls depend on model scale?
  •  Generalizable dense reward for long-horizon robotic tasks
  •  MUSS: Multilevel subset selection for relevance and diversity
  •  Bridging language models and knowledge graphs with controlled natural languages