Tencent's ANGLE unifies ad retrieval, relevance and ranking in a single LLM
_reachsumit · x · 2026-09-17
Tencent researchers published the ANGLE framework, replacing discrete semantic identifiers (SIDs) with hierarchical text representations for sponsored search ads.
Key points:
- Traditional multi-stage cascade architectures optimize modules independently, causing inconsistent objectives and premature elimination of good candidates; prior LLM-based methods rely on SIDs that need extensive SFT memorization, generalize poorly to unseen ads, and decode inefficiently.
- ANGLE represents ads as hierarchical text—high-level commercial intent plus fine-grained ad abstracts—and performs retrieval, relevance and ranking within a single LLM, leveraging its full capacity to assess commercial value for real-time one-step retrieval.
More from Research
- Microsoft researcher points out BERT was already called an LLM back in 2019 — JFPuget · 2026-09-20
- Have we seen an acceleration in discoveries? Cyber spikes, math rises, algorithms flat — soumitrashukla9 · 2026-09-20
- TovanaEngine: a local world model trained on 50K+ real SWE-bench coding-agent runs — Decent-Ad9950 · 2026-09-20
- After Chess and Go: Can Any AI Engine Actually Beat Humans at Scrabble? — zuilserip · 2026-09-20
- Textbook author: 99.9% accuracy can mean zero scientific discoveries — bravo_abad · 2026-09-20
- rasbt: Jev's Secret Sauce Is Data, Not the Algorithm — Laya Rival Falls Far Short — RichmanRonald · 2026-09-20