AI Search
Using the new decision models to rerank search results does not beat a purpose-tuned reranker, concludes this University of Passau paper.
AI Search
In LLM-based text retrieval, documents can be encoded offline, but queries arrive in real time, so every query runs the full LLM, which keeps throughput low and serving costly.
AI Search
In contrastive models like CLIP and CLAP, the two modalities are trained to align yet their embeddings often sit in offset regions, a phenomenon called the modality gap.
AI Search
Some multimodal embedding models first have the model write reasoning about the input and then embed it along with the input, producing richer vectors.
AI Search
Much enterprise knowledge sits in scanned PDFs that ordinary search cannot read.
Personalization
Local-services and live-commerce ads are ranked and billed per store or livestream room, but users actually buy one voucher or one product inside it, so directly predicting how much a click will sell gives a noisy label that is hard to learn.
Personalization
CTR models usually encode long-term and short-term behavior separately, then concatenate or gate them into one interest vector, so one-off noise in recent behavior, such as stocking up during a sale or buying for someone else, gets written straight into long-term preference.
Personalization
To save compute, long-history recommenders compress a user's thousands of actions into a small set of cached vectors.
Personalization
Recommenders train on negatives drawn from items a user never clicked.
Infra
LLM reinforcement learning usually generates answers with one inference engine and trains with another framework.
Infra
Retrieval-augmented generation stitches several retrieved documents into the context, and recomputing their KV caches from scratch every time is expensive, so some methods store each document's KV cache and concatenate them on use.
Infra
Ranking keyword search results with BM25 in Qdrant used to mean turning every document and query into a sparse vector and tracking the average document length yourself, which drifts as data changes.
Infra
In LLM reinforcement learning post-training, generating answers often takes over 70% of training time.
Generative
Recommender systems usually run a separate retrieval model for the home feed, search and related items, because each surface wants something different: saves on the home feed, outbound clicks to buy on search, and continued browsing on related items.
Generative
Generative recommenders that write the next item's Semantic ID code by code have a systematic bias.
Generative
When shopping online, people usually look at a broad category before picking a specific item.
Generative
In generative recommendation with Semantic IDs, training only tells the model which item came next, so it learns nothing about how likely the other items were given the same history.
Embodied AI
In flow-matching vision-language-action models, solver steps and execution horizon turn out to be coupled, and the best setting can reverse.