AI Search
Dense retrievers usually take hard negatives from top-ranked documents that are not labeled positive, but ranking high only shows a document resembles the query, not why it should be excluded.
AI Search
Most RAG systems still retrieve with BM25 or vector similarity, and evidence the retriever misses cannot be recovered by any generator.
AI Search
Search agents often send sub-queries of just a few keywords, so a sub-query-only reranker promotes documents that match the words but do not help the real question.
AI Search
LLM query rewriting usually applies one instruction to every query, yet the right direction for a rewrite depends on the corpus.
AI Search
If a team wants to replace its RAG reranker with a large model, this paper says to tune the existing reranker first.
Personalization
Most production ranking models squeeze the user's behavior sequence into one fixed vector before it reaches the feature-interaction module, so later interaction layers can no longer see which clicks or add-to-carts actually happened.
Personalization
There are three common ways to trim the input side of a CTR model: drop low-quality training samples, switch off useless features, and prune useless embedding dimensions.
Personalization
In ad systems, recalibrating a CTR model on the last few hours of fresh data is usually treated as a routine step that can only help.
Personalization
In ad auto-bidding, a PID (proportional-integral-derivative) controller keeps adjusting bids according to spend pacing, yet one set of control parameters has to serve hundreds of thousands of campaigns with very different budgets, durations and traffic.
Infra
During traffic spikes, each service degrades on its own, so every service survives while the product's quality can collapse.
Infra
A multimodal query first searches text vectors and then fetches the matching image vectors. Both nearest-neighbor searches are fast.
Infra
Replicas in a distributed vector database keep results available but do not make them trustworthy.
Infra
In long-context inference, the largest memory cost is often the per-layer KV cache, the keys and values attention stores for every past token.
Infra
Most of the gain from lossless KV cache compression is eaten by decompression.