Author

JD Retail Technology

Official platform of JD Retail Technology, delivering insightful R&D news and a deep look into the lives and work of technologists.

814

Articles

Likes

3.3k

Views

Comments

Latest from JD Retail Technology

100 recent articles max

JD Retail Technology

Jul 6, 2026 · Artificial Intelligence

GROLE: Instance-Level Expert Routing for Incremental Learning in Oxygen AIIC Models

The paper introduces GROLE, a two‑stage incremental‑learning framework that builds a frozen pool of task‑specific LoRA experts and trains a lightweight instance‑level selector via reinforcement‑learning‑based gradient‑free optimization with Dirichlet sampling, achieving superior stability‑plasticity trade‑offs and state‑of‑the‑art results on multiple CL benchmarks.

GROLELarge Language ModelsLoRA

0 likes · 14 min read

GROLE: Instance-Level Expert Routing for Incremental Learning in Oxygen AIIC Models

JD Retail Technology

Jun 29, 2026 · Artificial Intelligence

Uni-AdGen: Unified Autoregressive Model for Personalized Image‑Text Ad Generation (CVPR 2026)

Uni‑AdGen unifies image and text generation in a single autoregressive framework, introduces a coarse‑to‑fine preference module and foreground‑aware control, and demonstrates superior performance on the million‑scale PAd1M dataset with novel evaluation metrics for personalized advertising.

Autoregressive Modelevaluation metricslarge-scale dataset

0 likes · 15 min read

Uni-AdGen: Unified Autoregressive Model for Personalized Image‑Text Ad Generation (CVPR 2026)

JD Retail Technology

Jun 25, 2026 · Artificial Intelligence

JD Donates Oxygen xLLM Inference Engine to OpenAtom Foundation to Accelerate Domestic AI Infra

JD announced the donation of its self‑developed Oxygen xLLM large‑model inference engine to the OpenAtom Open Source Foundation, detailing its service‑engine decoupled architecture, performance breakthroughs, multi‑chip support, and early industrial validations that aim to foster a collaborative domestic AI infrastructure ecosystem.

AI inferenceDomestic AI ecosystemEngineering Intelligence

0 likes · 8 min read

JD Donates Oxygen xLLM Inference Engine to OpenAtom Foundation to Accelerate Domestic AI Infra

JD Retail Technology

Jun 23, 2026 · Artificial Intelligence

How RAD‑DPO Aligns Preferences in OxygenSearch Generative Retrieval (SIGIR 2026)

This article analyzes the challenges of generative retrieval for e‑commerce (shared SID prefixes, noisy implicit feedback, and probability squeezing) and presents RAD‑DPO, a robust adaptive denoising direct preference optimization that uses session‑level multi‑label contrast, token‑level gradient detachment, and dynamic reward weighting to improve both effectiveness and training efficiency.

Preference OptimizationRAD-DPOe-commerce search

0 likes · 18 min read

How RAD‑DPO Aligns Preferences in OxygenSearch Generative Retrieval (SIGIR 2026)

JD Retail Technology

Jun 15, 2026 · Artificial Intelligence

Sparse‑Up: Sparse‑Voxel 3D Geometry Generation Boosts Product Reconstruction

The paper presents Sparse‑Up, a sparse‑voxel based 3D geometry generation framework that introduces learnable surface‑anchored upsampling and view‑partition rendering to eliminate redundant points, dramatically reduce memory usage, and achieve high‑fidelity product reconstruction, with experimental results surpassing TRELLIS, Hunyuan3D and other SOTA methods.

3D Geometry GenerationAI3DICASSP 2026

0 likes · 13 min read

Sparse‑Up: Sparse‑Voxel 3D Geometry Generation Boosts Product Reconstruction

JD Retail Technology

Jun 8, 2026 · Mobile Development

Accelerating Taro Native Static Layout Rendering on HarmonyOS

The article analyzes severe scroll jank on low‑end HarmonyOS devices caused by Taro Native's heavyweight card page, identifies main‑thread overload in layout phases 1, 4 and 5, proposes static node‑tree layout with custom measurement interception and font‑measurement caching, and reports a frame‑rate boost from 43 fps to 57 fps (~32.5% improvement).

CustomNodeHarmonyOSNODE_LAYOUT_RECT

0 likes · 8 min read

Accelerating Taro Native Static Layout Rendering on HarmonyOS

JD Retail Technology

Jun 2, 2026 · Artificial Intelligence

RTPrune: Two‑Stage Reading‑Inspired Token Pruning for Efficient DeepSeek‑OCR Inference

The paper presents RTPrune, a token‑pruning technique for DeepSeek‑OCR that exploits a two‑stage reading behavior in LLM decoding, first keeping high‑norm visual tokens and then fusing the rest via optimal‑transport matching with a dynamic pruning‑rate strategy, achieving up to 15% GFLOPs reduction and 18.9% speedup while preserving over 99% OCR accuracy across multiple benchmarks.

DeepSeek-OCROCR efficiencydynamic pruning

0 likes · 9 min read

RTPrune: Two‑Stage Reading‑Inspired Token Pruning for Efficient DeepSeek‑OCR Inference

JD Retail Technology

May 25, 2026 · Artificial Intelligence

How Adaptive Semantic IDs Enable Precise and Generalizable Generative Retrieval

The article introduces the SA²CRQ framework, which adaptively allocates semantic ID length and transfers residual knowledge to resolve head‑item ID collisions and tail‑item generalization gaps in large‑scale e‑commerce generative retrieval, achieving stable gains on both industrial and public datasets.

Adaptive QuantizationEmbeddingLong-tail Distribution

0 likes · 19 min read

How Adaptive Semantic IDs Enable Precise and Generalizable Generative Retrieval

JD Retail Technology

May 19, 2026 · Artificial Intelligence

Spectral Disentanglement and Enhancement: Teaching Multimodal Models to Denoise and Purify

The paper introduces the Spectral Disentanglement and Enhancement (SDE) framework, which uses singular value decomposition to separate strong semantic signals, weak auxiliary signals, and noise, applies curriculum‑based spectral enhancement, and jointly optimizes a dual‑domain contrastive loss, achieving markedly improved robustness and generalization on large‑scale multimodal benchmarks.

contrastive learningdual-domain lossmultimodal learning

0 likes · 14 min read

Spectral Disentanglement and Enhancement: Teaching Multimodal Models to Denoise and Purify

JD Retail Technology

Apr 28, 2026 · Mobile Development

How Taro 5.0 Delivers a High‑Performance iOS Rendering Layer and a Unified Cross‑Platform UI Framework

The article explains why existing cross‑platform solutions cannot meet Taro’s demands for high performance, stability, full Taro ecosystem compatibility and low integration cost, and describes the design of a dual‑thread iOS rendering layer, a mixed native‑view + layer approach, and the C++‑based TaroUI framework that provides consistent, high‑speed UI components, a rich‑text engine, virtual list widgets and an extensible architecture for future platforms.

C++ frameworkCross‑Platform UITaRO

0 likes · 19 min read

How Taro 5.0 Delivers a High‑Performance iOS Rendering Layer and a Unified Cross‑Platform UI Framework