Author

Alimama Tech

Official Alimama tech channel, showcasing all of Alimama's technical innovations.

255

Articles

Likes

1.9k

Views

Comments

Latest from Alimama Tech

100 recent articles max

Alimama Tech

Oct 22, 2026 · Artificial Intelligence

Data Stays, Model Moves: SDH PaaS 3.0's Federated Learning Productization

SDH PaaS 3.0 integrates vertical federated learning into data collaboration workflows, enabling platform and business parties to jointly train conversion prediction models without sharing raw data, covering role assignment, sample alignment, encrypted intermediate exchange, privacy controls, task orchestration with approvals, and engineering optimizations for large-scale stability.

SDH PaaS 3.0conversion predictiondata collaboration

0 likes · 15 min read

Data Stays, Model Moves: SDH PaaS 3.0's Federated Learning Productization

Alimama Tech

Sep 16, 2026 · Information Security

Alibaba Mama's SDH Platform Secures Dual ISO 27001/27701 Certification for Privacy Computing Governance

Alibaba Mama's Secure Data Hub (SDH) privacy computing PaaS platform has achieved ISO/IEC 27001 and ISO/IEC 27701 dual certification, establishing a governance framework that maps security and privacy standards to platform mechanisms for asset permission management, operation control, traceability, risk governance, and auditability across multi-party data collaboration.

Alibaba MamaGovernance FrameworkISO 27001

0 likes · 14 min read

Alibaba Mama's SDH Platform Secures Dual ISO 27001/27701 Certification for Privacy Computing Governance

Alimama Tech

Sep 3, 2026 · Artificial Intelligence

MOON 3.0: Attribute Reasoning for E-commerce Multimodal Representation (ACM MM'26)

MOON 3.0 shifts e-commerce multimodal representation from direct encoding to explicit attribute reasoning, using Multi-head Modality Fusion, joint contrastive and reinforcement learning, and fine-grained residual enhancement to improve fine-grained retrieval and classification across multiple benchmarks.

Contrastive LearningE-commerceMLLM

0 likes · 24 min read

MOON 3.0: Attribute Reasoning for E-commerce Multimodal Representation (ACM MM'26)

Alimama Tech

Aug 20, 2026 · Artificial Intelligence

TransVAE: Combining CNN Local Bias with Transformer Global Modeling for Scalable Visual Tokenizers

TransVAE merges early‑stage CNN local feature extraction with deep‑layer Transformer global reasoning, uses RoPE‑only positional encoding and ConvFFN to enable resolution‑agnostic scaling, and demonstrates stable improvements in reconstruction quality, latent alignment, and downstream generation across model sizes from 44 M to 2.3 B parameters.

CNNScalingTransVAE

0 likes · 12 min read

TransVAE: Combining CNN Local Bias with Transformer Global Modeling for Scalable Visual Tokenizers

Alimama Tech

Jul 8, 2026 · Artificial Intelligence

Optimizing Qwen Series Inference with SGLang‑ATOM on AMD Instinct GPUs

Alibaba Mama AI Infra and AMD collaborated to integrate SGLang‑ATOM plugin mode for Qwen 3.5/3.6 models on MI308X GPUs, detailing multimodal, GDN, KV‑cache, and parallel‑strategy optimizations that achieve low‑latency, high‑throughput inference suitable for complex advertising workloads.

AMD ATOMGPU optimizationLLM inference

0 likes · 18 min read

Optimizing Qwen Series Inference with SGLang‑ATOM on AMD Instinct GPUs

Alimama Tech

Jun 4, 2026 · Artificial Intelligence

ICML 2026 Highlights: Five Taotian Group Papers Pushing Multimodal AI Boundaries

The article showcases five ICML 2026 papers from the Taotian Group that tackle core multimodal AI challenges—interactive video try‑on, high‑resolution vision, e‑commerce video reasoning, sparse‑reward reinforcement learning, and curriculum learning for large language models—detailing their problem statements, novel solutions, and strong experimental results.

ICML 2026Large Language Modelsbenchmark

0 likes · 15 min read

ICML 2026 Highlights: Five Taotian Group Papers Pushing Multimodal AI Boundaries

Alimama Tech

May 28, 2026 · Artificial Intelligence

13 KDD'26 Papers from Taobao: Scaling Laws, World Models and New AI Paradigms

The article highlights thirteen Taobao‑group papers accepted at KDD 2026, covering large‑model scaling laws, end‑to‑end generative recommendation, CTR prediction, interactive recommendation agents, LLM‑based pricing, robust auto‑bidding, two‑stage auctions, generative world models, multi‑attribution conversion, uplift modeling and long‑term causal estimation for e‑commerce systems.

CTR predictionGenerative ModelsKDD 2026

0 likes · 29 min read

13 KDD'26 Papers from Taobao: Scaling Laws, World Models and New AI Paradigms

Alimama Tech

May 28, 2026 · Artificial Intelligence

TAR: Multi‑Scale Trajectory Model Fixes Granularity Mismatch, Raising CTR >12%

The paper introduces the Trajectory Auto‑Regressive (TAR) model, which uses multi‑scale trajectory generation, a VQ‑VAE latent compression, and a state‑action fusion architecture to address granularity mismatch between fine‑grained decision steps and coarse‑grained feedback in online advertising, achieving over 12% CTR lift, smoother budget pacing, and faster inference compared to prior baselines.

Budget PacingMulti-Scale GenerationOnline Advertising

0 likes · 18 min read

TAR: Multi‑Scale Trajectory Model Fixes Granularity Mismatch, Raising CTR >12%

Alimama Tech

May 21, 2026 · Artificial Intelligence

Bridging LLMs' Social Gap: Graphia Uses Social Graphs as Supervision for Full Macro‑Micro Alignment

Graphia, a new LLM‑based social simulation framework, leverages social graph data as high‑quality supervision to jointly align microscopic interaction predictions and macroscopic network structures, achieving significant gains on TDGG and IDGG benchmarks across three real‑world datasets.

Evaluation MetricsGraphiaLLM

0 likes · 12 min read

Bridging LLMs' Social Gap: Graphia Uses Social Graphs as Supervision for Full Macro‑Micro Alignment

Alimama Tech

May 14, 2026 · Artificial Intelligence

How LLM-Auction Lets Large Language Models Learn to Auction Marketing Content Within Answers

The article presents LLM-Auction, a novel AI‑native marketing mechanism that unifies ad allocation and answer generation by training large language models to conduct auctions directly on their output distribution, achieving higher allocation efficiency without extra inference cost.

AI-native advertisingLLM-AuctionLarge Language Models

0 likes · 17 min read

How LLM-Auction Lets Large Language Models Learn to Auction Marketing Content Within Answers