vipin wiki

LLM Recommendation Research Map

2 min read

LLM Recommendation Research Map

Summary

Vipin currently has three local research collections that together form an initial literature base for LLM + recommendation work.

Current Collections

  • NH baseline set:
    • 21 PDFs
    • appears to focus on baseline and recent papers for LLM-based recommendation, especially sequential and controllable settings
  • NR baseline set:
    • 14 PDFs
    • appears to be a second curated baseline subset within the Uncertainty-LLM4Rec project
  • broader recommendation library:
    • 28 top-level PDFs and 15 topic/project subfolders
    • appears to span the broader literature landscape

Emerging Topic Clusters

  • sequential recommendation
  • explainable recommendation
  • conversational recommendation
  • controllable recommendation
  • multimodal recommendation
  • knowledge-graph and retrieval augmentation
  • bias, fairness, and job recommendation
  • efficiency, distillation, and deployment trade-offs

Likely Reading Strategy

  1. Identify the papers that are true baselines for Vipin’s current project.
  2. Separate methodological baselines from adjacent inspiration papers.
  3. Create per-paper notes for the most central works.
  4. Build one comparison note per major cluster, especially:
    • sequential recommendation
    • explainability
    • controllability
    • robustness and uncertainty

Important Unknowns

  • What do NH and NR stand for in the project taxonomy?
  • Which subset is required reading for the current paper or experiment?
  • Which papers are already deeply read versus merely collected?

Counterpoints and Gaps

  • the map is collection-driven rather than paper-by-paper, so it is still coarse and may overstate coverage in some clusters
  • folder presence does not imply equal relevance to Vipin’s active research question
  • uncertainty-specific structure is still missing, so the current map is a literature inventory, not yet a working thesis
  1. Pick 3 to 5 cornerstone papers from these collections.
  2. Ingest those papers one by one into dedicated source notes.
  3. Create a comparison page for the baselines used in Uncertainty-LLM4Rec.
  4. Add a page specifically for uncertainty in LLM-based recommendation once the key papers are identified.

Graph View

Backlinks