Author

Alibaba International Intelligent Technology

Alibaba International Tech – Official channel of the Intelligent Technology team, sharing cutting‑edge AI applications and innovations in Alibaba's global e‑commerce business.

Articles

Likes

167

Views

Comments

Latest from Alibaba International Intelligent Technology

30 recent articles

Alibaba International Intelligent Technology

Oct 7, 2026 · Artificial Intelligence

GRC: Self-Diagnosis and Correction for Generative Recommendation (CIKM'26 Oral)

This paper introduces GRC, a generation-reflection-correction framework for generative recommendation that enables models to self-diagnose and correct full draft semantic ID sequences via structured reflection labels, GRPO reinforcement learning, and entropy-guided scheduling, achieving state-of-the-art results on public and industrial datasets with minimal latency overhead.

CIKM 2026GRPOReflection

0 likes · 12 min read

GRC: Self-Diagnosis and Correction for Generative Recommendation (CIKM'26 Oral)

Alibaba International Intelligent Technology

Oct 1, 2026 · Artificial Intelligence

Cross-Country Code-Mixing Feeds Big-Market Data to Small Markets in Generative Recommendation

The paper proposes CMRec, a cross-country code-mixing method for generative recommendation that constructs a unified semantic codebook and mixes token-level interactions across countries with dual semantic and behavioral constraints, achieving significant gains on public and industrial datasets and a 1.77% ad revenue and 2.64% order increase in online A/B tests at Lazada.

CIKM 2026Code-MixingCross-Country Recommendation

0 likes · 10 min read

Cross-Country Code-Mixing Feeds Big-Market Data to Small Markets in Generative Recommendation

Alibaba International Intelligent Technology

Sep 18, 2026 · Artificial Intelligence

UTTSI: Test-Time Selective Inference Boosts CTR Prediction Without Retraining

Alibaba researchers propose UTTSI, a training-free, model-agnostic framework that estimates per-sample uncertainty at inference time using frequency priors and gradient-based attribution, then adaptively allocates compute — filtering noisy features and exploring multiple inference paths only for uncertain samples — achieving a 5.3% CTR lift in a 7-day online A/B test with just 2.8× average model calls.

CIKM 2026CTR predictionCount-Min Sketch

0 likes · 12 min read

UTTSI: Test-Time Selective Inference Boosts CTR Prediction Without Retraining

Alibaba International Intelligent Technology

Sep 3, 2026 · Artificial Intelligence

Masked Diffusion Challenges Left-to-Right Decoding in Generative Recommendation (Recsys'26)

This paper introduces MDGR, a masked diffusion framework for generative recommendation that replaces autoregressive left-to-right decoding with a parallel mask-denoising process, achieving up to 6.56% relative improvement offline and significant online gains in Alibaba's advertising platform.

AlibabaMasked DiffusionOnline A/B Testing

0 likes · 17 min read

Masked Diffusion Challenges Left-to-Right Decoding in Generative Recommendation (Recsys'26)

Alibaba International Intelligent Technology

Aug 21, 2026 · Artificial Intelligence

Alibaba's 10 CIKM 2026 Papers: Multimodal Large Models Transform E‑Commerce Search & Recommendation

Alibaba International Intelligent Technology presented ten CIKM 2026 papers that introduce multimodal large‑model techniques—such as SAM‑D2Q, C2P, PRO‑Bid, UTTSI, GRC, CDNet, CAIM and SORT—to overcome bottlenecks in e‑commerce search, recommendation, auto‑bidding and CTR prediction, delivering substantial offline metric lifts and significant online gains in CTR, CVR, GMV and revenue.

CTR predictionLarge Language Modelsattribute extraction

0 likes · 22 min read

Alibaba's 10 CIKM 2026 Papers: Multimodal Large Models Transform E‑Commerce Search & Recommendation

Alibaba International Intelligent Technology

Jul 30, 2026 · Artificial Intelligence

FedAFD Enables Strong Cloud and Edge Performance in Multimodal Federated Learning

FedAFD, a CVPR 2026 paper, combines bi‑level adversarial alignment, granularity‑aware feature fusion, and similarity‑guided ensemble distillation to simultaneously boost cloud‑side global models and edge‑side personalization in multimodal federated learning, achieving state‑of‑the‑art results with only 20 communication rounds.

Adversarial AlignmentCVPR 2026Ensemble Distillation

0 likes · 11 min read

FedAFD Enables Strong Cloud and Edge Performance in Multimodal Federated Learning

Alibaba International Intelligent Technology

Jul 3, 2026 · Artificial Intelligence

How Generative Pretraining Overcomes Discriminative Model Bottlenecks in Ad Ranking

The article presents UserLLM, a three‑stage framework—generative pre‑training, discriminative SFT, and CTR fusion—that leverages ultra‑long user behavior sequences to address the compression and scaling limits of traditional discriminative ranking models, demonstrating significant offline gains and validated scaling‑law behavior across token, layer, and width dimensions.

Large Language ModelsScaling LawUserLLM

0 likes · 17 min read

How Generative Pretraining Overcomes Discriminative Model Bottlenecks in Ad Ranking

Alibaba International Intelligent Technology

Jun 26, 2026 · Artificial Intelligence

How CCVTON Achieves SOTA Virtual Try-On Without Large-Scale Paired Data

CCVTON introduces a cycle‑consistent diffusion framework that trains on massive unpaired single‑model images, uses a two‑stage garment‑aware mask generation and multi‑criterion filtering to overcome paired‑data scarcity, and reaches state‑of‑the‑art results on VITON‑HD and DressCode.

CCVTONCycle-ConsistencyVirtual Try-On

0 likes · 11 min read

How CCVTON Achieves SOTA Virtual Try-On Without Large-Scale Paired Data

Alibaba International Intelligent Technology

Jun 18, 2026 · Artificial Intelligence

GeoReward: Enabling Vision‑Language Models to Sense Market Context for Country‑Specific Ad Creative

The paper identifies Contextual Variable Overestimation in vision‑language models, introduces the MACP multi‑country ad preference dataset, proposes the three‑gate GeoReward framework to restore sensitivity to sparse country cues, and demonstrates superior accuracy, sensitivity, and controllable ad generation across ten markets.

Cross-Market PreferenceDataset MACPGeoReward

0 likes · 19 min read

GeoReward: Enabling Vision‑Language Models to Sense Market Context for Country‑Specific Ad Creative

Alibaba International Intelligent Technology

Jun 12, 2026 · Artificial Intelligence

Personalized World Knowledge Lets Large Models Truly Understand Users in Generative Recommendation

This article introduces LWGR, a framework that uses personalized soft prompts generated by parallel codebooks and Lagrangian‑constrained knowledge fusion to integrate large language model world knowledge into generative recommendation, overcoming fixed‑prompt limitations and knowledge‑behavior conflicts, and demonstrates superior performance on public and industrial datasets with notable revenue gains in online A/B tests.

Knowledge FusionLagrangian constraintgenerative recommendation

0 likes · 11 min read

Personalized World Knowledge Lets Large Models Truly Understand Users in Generative Recommendation