Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,695 papers · 148 categories

Trend · papers per month

count% of corpus
0.3%0.5%0.8%0.1% · Aug 202219922001200920172026
6 results for “Two-Tower”
Practice with Graph-based ANN Algorithms on Sparse Data: Chi-square Two-tower model, HNSW, Sign Cauchy Projectionscs.IR

Efficient ANN search for sparse embeddings in ads targeting.

problem Efficiently searching near neighbors in sparse data for applications like ads targeting.
method Graph-based ANN algorithms (HNSW, chi-square two-tower model, Sign Cauchy Projections).
result Sparse embeddings and ANN algorithms improve efficiency in EBR applications.
sparse dataANN searchgraph-based algorithmschi-squaresign cauchy projections
2023-06-13abs ↗pdf ↗
Theoretical Guarantees for LT-TTD: A Unified Transformer-based Architecture for Two-Level Ranking Systemscs.IR

Unified transformer-based LT-TTD improves ranking efficiency and quality.

problem Decoupled L1 and L2 models in recommendation and search systems cause irreversible error propagation and suboptimal ranking.
method LT-TTD combines two-tower models with transformer expressivity in a unified listwise learning framework, providing theoretical guarantees and UPQE evaluation.
result LT-TTD reduces irretrievable relevant items and achieves better global optimization than disjoint training.
TransformerUnified ArchitectureTwo-TowerRanking SystemsListwise Learning
2025-05-07abs ↗pdf ↗
Towards User Friendly Medication Mapping Using Entity-Boosted Two-Tower Neural Networkcs.CL

Improves medication name inference for telemedicine and conversational agents.

problem Challenges in mapping user-friendly medication names to standardized ones.
method Entity-boosted two-tower neural network for ranking SMN to DMP.
result State-of-the-art results achieved with improved attention-based ranking.
medication inferencetelemedicineconversational agentsentity linkingtwo-tower neural network
2020-06-17abs ↗pdf ↗
MuLan: A Joint Embedding of Music Audio and Natural Languageeess.AS

MuLan links music audio to natural language tags.

problem Traditional music tagging systems use rigid attributes; MuLan aims to link audio directly to natural language.
method Joint audio-text embedding model trained on 44 million music recordings and text annotations.
result MuLan's embeddings enable zero-shot functionalities and transfer learning.
music taggingnatural languageaudio-text embeddingzero-shottransfer learning
2022-08-26abs ↗pdf ↗
Modeling Content Creator Incentives on Algorithm-Curated Platformscs.GT

Modeling incentives for content creators on algorithm-curated platforms.

problem Maximizing exposure for content creators on algorithmic platforms.
method Formalized exposure game model, proving effects of algorithmic choices on equilibria, proposing tools for finding equilibria.
result Algorithmic choices significantly affect content exposure and creator behavior.
content creatorsalgorithmic platformsexposure gameNash equilibriacreator behavior
2022-06-27abs ↗pdf ↗
Virtual an arrow Temperley--Lieb algebras, Markov traces, and virtual link invariantsmath.GR

Defines new algebras for virtual link invariants, matching known polynomials.

problem Developing new mathematical structures for virtual link invariants.
method Introducing two towers of algebras, VTL and ATL, and determining their presentations and Markov traces.
result The invariants derived from the Markov traces match known polynomials for virtual links.
virtual linkinvariantsTemperley-LiebMarkov tracesalgebras
2020-09-17abs ↗pdf ↗