Author

Xiaohongshu Tech REDtech

Official account of the Xiaohongshu tech team, sharing tech innovations and problem insights, advancing together.

115

Articles

Likes

324

Views

Comments

Latest from Xiaohongshu Tech REDtech

100 recent articles max

Xiaohongshu Tech REDtech

Aug 23, 2024 · Artificial Intelligence

Xiaohongshu REDtech Live: Presentation of Recent Top‑Conference Papers (Recruitment Session)

On August 24, 2024, Xiaohongshu’s technical team will livestream a four‑hour REDtech session across WeChat Channels, its recruitment account, and Bilibili, showcasing recent top‑conference papers—from ACL and CVPR to ICLR and AAAI—covering innovations such as KV‑cache compression, zero‑shot image generation, early‑stopping self‑consistency, negative‑sample‑aware distillation, and real‑time nearest‑neighbor search, while allowing live interaction and offering surprise merchandise.

AIConference PapersXiaohongshu

0 likes · 18 min read

Xiaohongshu REDtech Live: Presentation of Recent Top‑Conference Papers (Recruitment Session)

Xiaohongshu Tech REDtech

Aug 1, 2024 · Artificial Intelligence

Xiaohongshu Search Advertising Recall: Practices, Metrics, and Large‑Model Integration

Xiaohongshu’s search advertising recall system evolves from keyword bidding to BERT‑based vector retrieval and LLM‑enhanced query rewriting, using dual semantic and efficiency models, water‑level metrics, and GPU‑accelerated engineering to achieve 80 % click coverage, 60 % conversion coverage and a 5 % CPM lift.

Artificial IntelligenceSemantic Retrievalefficiency optimization

0 likes · 33 min read

Xiaohongshu Search Advertising Recall: Practices, Metrics, and Large‑Model Integration

Xiaohongshu Tech REDtech

Jul 29, 2024 · Artificial Intelligence

Scaling Laws for Dense Retrieval: Empirical Study of Model Size, Training Data, and Annotation Quality

The award‑winning study shows that dense retrieval performance follows precise power‑law scaling with model size, training data quantity, and annotation quality, introduces contrast entropy for evaluation, validates joint scaling formulas on MS MARCO and T2Ranking, and uses cost models to guide budget‑optimal resource allocation.

Information RetrievalModel Sizeannotation quality

0 likes · 13 min read

Scaling Laws for Dense Retrieval: Empirical Study of Model Size, Training Data, and Annotation Quality

Xiaohongshu Tech REDtech

Jun 24, 2024 · Databases

Distributed Parallel Multi‑Hop Query Optimization in REDgraph Graph Database

To meet Xiaohongshu’s strict P99 latency targets for multi‑hop graph queries, the REDgraph team built a distributed parallel execution framework that eliminates global barriers, pushes operators to storage nodes, uses routing operators and a NeighborCache, cutting three‑hop query latency by over 50 % and enabling production deployment.

REDgraphdistributed parallelgraph database

0 likes · 29 min read

Distributed Parallel Multi‑Hop Query Optimization in REDgraph Graph Database

Xiaohongshu Tech REDtech

Jun 20, 2024 · Artificial Intelligence

Xiaohongshu 2024 Large Model Frontier Paper Sharing Live Event

On June 27, 2024, Xiaohongshu’s technical team will livestream a two‑hour session across WeChat Channels, Bilibili, Douyin and Xiaohongshu, showcasing six top‑conference papers on large‑model advances—including early‑stopping and fine‑grained self‑consistency, novel evaluation methods, negative‑sample‑assisted distillation, and LLM‑based note recommendation—followed by a Q&A and recruitment briefing.

AI researchKnowledge DistillationSelf-Consistency

0 likes · 12 min read

Xiaohongshu 2024 Large Model Frontier Paper Sharing Live Event

Xiaohongshu Tech REDtech

Jun 17, 2024 · Artificial Intelligence

Xiaohongshu Audio-Video Architecture Team Wins Top Awards in CVPR NTIRE 2024 Challenges

Xiaohongshu’s audio‑video architecture team secured second place in the RAIM challenge and first in the S‑UGC VQA challenge at CVPR NTIRE 2024 by improving generative image restoration with SUPIR, DeSRA and a Fusion model, and enhancing video quality assessment using LIQE, Q‑Align and FAST‑VQA, then deploying these methods for live‑stream denoising, intelligent transcoding and cloud‑based super‑resolution, achieving high PLCC/SROCC scores and up to 33 % bandwidth savings.

AICVPR NTIRE 2024Denoising

0 likes · 25 min read

Xiaohongshu Audio-Video Architecture Team Wins Top Awards in CVPR NTIRE 2024 Challenges

Xiaohongshu Tech REDtech

May 31, 2024 · Frontend Development

REDtech Salon: Frontier Exploration of WebGPU and Interactive Rendering Technologies

The REDtech Salon in Shanghai showcased WebGPU and interactive rendering advances, with W3C’s François Daoust demonstrating WebCodecs, Streams, and GPU‑accelerated video processing, while Xiaohongshu’s Huang Yuejia and Wang Ning highlighted business‑driving projects, migration from WebGL to WebGPU, and future AR‑enabled frontend innovations.

Interactive RenderingStreams APIW3C

0 likes · 8 min read

REDtech Salon: Frontier Exploration of WebGPU and Interactive Rendering Technologies

Xiaohongshu Tech REDtech

May 23, 2024 · Cloud Native

Cloud-Native Architecture and Tiered Storage for Xiaohongshu Kafka: Cost Reduction, Elastic Migration, and Performance Optimization

Xiaohongshu's big-data storage team built cloud-native architecture with tiered storage, containerized Kafka, and custom load balancer, cutting storage costs up to 60%, enabling minute‑level elastic migration, improving scaling efficiency tenfold, and boosting performance via caching and batch reads.

KafkaTiered Storagecost optimization

0 likes · 20 min read

Cloud-Native Architecture and Tiered Storage for Xiaohongshu Kafka: Cost Reduction, Elastic Migration, and Performance Optimization

Xiaohongshu Tech REDtech

May 10, 2024 · Artificial Intelligence

SIF3D: Sense‑Informed Forecasting of 3D Human Motion with Multimodal Attention

SIF3D is a scene‑aware 3D human motion forecasting framework that fuses observed motion, 3D point‑cloud scenes, and gaze through novel ternary intention‑aware and semantic‑coherence‑aware attention mechanisms, encoding with PointNet++ and Transformers, and decoding with a graph‑convolutional network, achieving state‑of‑the‑art results on GIMO and GTA‑1M benchmarks.

3D scene understandingCVPR2024SIF3D

0 likes · 15 min read

SIF3D: Sense‑Informed Forecasting of 3D Human Motion with Multimodal Attention

Xiaohongshu Tech REDtech

Apr 28, 2024 · Artificial Intelligence

Generative Dense Retrieval: Memory Can Be a Burden

The paper introduces Generative Dense Retrieval (GDR), a two‑stage retrieval framework that first maps queries to memory‑efficient document‑cluster identifiers and then uses dense vectors to locate individual documents, achieving higher recall and better scalability than traditional generative retrieval while incurring modest latency and capacity trade‑offs.

Information RetrievalMemory Mechanismgenerative dense retrieval

0 likes · 13 min read

Generative Dense Retrieval: Memory Can Be a Burden