Tagged articles
Multimodal Data
38 articles · Page 1 of 1
Qunce Intelligence processes 300K hours of embodied intelligence data (10TB/day) using MaxCompute Serverless and MaxFrame, enabling video decoding, alignment, model inference, and quality checks in a unified pipeline without self-managed peak clusters, handling PB-scale multimodal sensor data with fault tolerance.
Lance FormatMaxComputeMaxFrame
0 likes · 15 min read

Experts at the Triple T summit debate whether embodied AI pre-training should prioritize massive human video datasets or real-robot interaction data, discussing tactile sensing, data recipes, evaluation methods, and cost reduction strategies for scaling embodied intelligence.
Multimodal Datacontinuous learningdata recipes
0 likes · 14 min read

This article introduces Daft, Ray, and Lance as the core layers of modern AI data infrastructure: Daft handles multimodal data computation, Ray manages distributed scheduling across heterogeneous resources, and Lance provides a lakehouse format optimized for vector search and random access, together replacing fragmented toolchains.
AI data infrastructureDaftLance
0 likes · 11 min read

At Yunqi 2026, Alibaba Cloud unveiled its Agentic Lake + Agentic Streaming architecture, combining DLF, Flink, EMR Ray/Daft, StarRocks, and Milvus to provide a unified multimodal data infrastructure for AI agents in autonomous driving, embodied intelligence, and enterprise operations.
Agentic LakeAgentic StreamingAlibaba Cloud
0 likes · 15 min read

StarRocks now fully supports Paimon 2.0, integrating columnar analytics with vector, full-text, and scalar indexes into a single SQL execution framework for multimodal lakehouse workloads, enabling unified data management, retrieval, and relational analysis on object storage without data movement.
Data EvolutionGlobal IndexMultimodal Data
0 likes · 34 min read

At Yunqi 2026, Alibaba Cloud unveiled a strategic upgrade to its ODPS big data platform, introducing agent-native architecture and multimodal computing across MaxCompute, Hologres, and DataWorks to support large model data refinement, embodied intelligence, and autonomous data operations with SQL AI, vector search, data sandboxes, and semantic knowledge graphs.
Agentic AIData SandboxDataWorks
0 likes · 21 min read

StarRocks introduces multimodal capabilities to handle structured, semi-structured, and unstructured data in a single SQL pipeline, combining object processing, full-text and vector search, and lakehouse integration with Paimon and Lance for unified retrieval and analysis.
AI analyticsHybrid RetrievalLance
0 likes · 24 min read

Volcano Engine's LAS video fine-grained understanding operator enables automated, scalable video data pipelines for embodied intelligence and autonomous driving, achieving 94.41% dual-arm accuracy, 100k hours/day processing, 90%+ data usability, and cutting annotation effort by 200+ person-days daily while reducing iteration cycles from weeks to days.
Autonomous DrivingLASMultimodal Data
0 likes · 10 min read

At AICon 2026, Ma Jin explained how Lance’s layered lakehouse format—spanning file, table, index and catalog layers—addresses AI workloads by supporting wide columns, cheap schema evolution, mixed storage and low‑latency random reads, enabling unified handling of training data, RAG retrieval and Agent memory.
AI-native FormatAgent MemoryLance
0 likes · 16 min read

The paper proposes a five‑layer Data Pyramid for embodied manipulation, analyzing each layer—real‑robot, UMI‑style, ego‑centric, simulation, and generic vision‑language data—through quality, diversity, reusability and physical fidelity dimensions, and discusses how these layers can be combined for training embodied AI models.
Data PyramidEmbodied AIMultimodal Data
0 likes · 15 min read

The article argues that post-DCMM 2.0, data governance must evolve from asset management to building high-quality datasets—trustworthy, semantically clear, quality-measurable, version-traceable, and compliant—to reliably support AI training, evaluation, knowledge augmentation, and agent workflows, requiring semantic foundations and AI data engineering.
AI data engineeringDCMM 2.0Multimodal Data
0 likes · 12 min read

The article surveys the leading embodied‑intelligence datasets—including ALOHA, Open X‑Embodiment, AGIBOT WORLD, RH20T and RoboMIND—detailing their technical features, scale, multimodal content, and how they support model training, benchmark evaluation, and practical robot development pipelines.
ALOHAMultimodal DataOpen X-Embodiment
0 likes · 14 min read


ITPUBJul 2, 2026 · Databases
Lakebase, the core engine of OceanBase AI Database, unifies storage, management, computation and search for structured, unstructured and vector data, offering multimodal data capabilities, flexible deployment modes, and AI‑oriented use cases such as autonomous driving and financial document analytics.
AI DatabaseAI applicationsData Management
0 likes · 9 min read

The article argues that as AI agents replace human users, the real bottleneck for enterprise AI shifts from model performance to data infrastructure, and explains how OceanBase’s AI‑native database—Lakebase—addresses multimodal data, hybrid search, agent safety, and massive logical tables to enable production‑grade AI applications.
AI DatabaseAgent-friendlyData Infrastructure
0 likes · 16 min read

Loop Engineering shifts AI from single prompts to continuous feedback loops, and by capturing human perception‑decision‑action‑feedback cycles with multimodal signals, the new Ego‑NeuroLoop paradigm promises far more data‑efficient embodied intelligence than existing ego‑centric video datasets.
Ego-NeuroLoopEmbodied AILoop Engineering
0 likes · 11 min read

The Flink Forward Asia 2026 conference in Shenzhen showcased Apache Flink's evolution to Agentic Streaming for AI, introduced the multimodal Agentic Lake built on Apache Paimon 2.0, announced Fluss 1.0 as a real‑time context layer, and highlighted performance gains over competing stacks such as Ray and Daft.
Agentic StreamingApache FlinkApache Fluss
0 likes · 13 min read

The article analyzes how the AI era shifts data from a passive by‑product to a core driver of large‑model performance, traces the evolution of data products from the DBA era through big‑data to AI‑native solutions, and details Volcano Engine’s four‑layer AI data platform that closes the data‑to‑model‑to‑Agent loop.
AIAgentMultimodal Data
0 likes · 12 min read

The article analyzes how enterprise‑level AI agents, when deployed in production, struggle with factual data retrieval despite semantic capabilities, and argues that real‑time, low‑latency, multimodal analytics—exemplified by systems like Apache Doris and SelectDB—are the essential data entry points for successful Agent deployments.
AI agentApache DorisMultimodal Data
0 likes · 9 min read

The article examines how Alibaba Cloud OSS Vector Bucket solves the data‑scattered, costly, and inefficient retrieval problems of massive multimodal media asset platforms by unifying storage, providing semantic vector search, and cutting operational expenses up to 95%.
Multimodal DataOSS Vector BucketVector Search
0 likes · 9 min read

This article details Hisense's journey to create an AI‑ready multimodal data platform, covering the challenges of integrating diverse business systems, the shift from a Hadoop‑based architecture to a cloud‑native data lake, the JuData governance and development platform, and six practical scenarios that demonstrate unified ingestion, metadata management, rule‑based quality control, intelligent asset retrieval, and future AI‑driven DataOps capabilities.
AI PlatformDataOpsJuData
0 likes · 23 min read

Lance File Format v2.2 tackles the AI data explosion by delivering hundred‑fold random‑read performance, advanced two‑layer compression, zero‑cost schema evolution, Git‑style versioning, external blob handling, and a roadmap toward native media support and intelligent encoding, positioning it as a core infrastructure for large‑scale multimodal workloads.
CompressionIO performanceLance
0 likes · 14 min read

The article explains how Alibaba Cloud's MaxCompute has been transformed into a cloud‑native Data+AI platform, detailing its layered architecture, multimodal storage, model management, hybrid compute scheduling, SQL AI functions, the MaxFrame Python framework, and several enterprise case studies that demonstrate performance gains and flexible resource orchestration.
AI IntegrationData+AIMaxCompute
0 likes · 11 min read

In the AI era, data consumption moves from a few technical users to all business staff, forcing a fundamental redesign of data governance across five dimensions—resource consumption, frequency, semantics, knowledge base, and modality—and proposing three actionable strategies to make data semantically rich, fully multimodal, and AI‑consumable.
AIMultimodal Datadata governance
0 likes · 18 min read

Paxini, together with JD Cloud, Tencent Cloud, and Baidu Intelligent Cloud, launches the world’s first hundred‑billion‑scale, full‑modal, high‑degree‑of‑freedom embodied AI data cloud mall, offering instant online data procurement, end‑to‑end model training pipelines, and validated performance gains in both lab and real‑world robot tasks.
Embodied AIModel TrainingMultimodal Data
0 likes · 13 min read

This article analyzes how the Daft‑Ray‑Lance stack tackles the challenges of multimodal AI workloads by offering a high‑performance Rust engine, adaptive back‑pressure, seamless Ray‑based distributed scheduling, and a storage format optimized for random access, vector indexing, and zero‑copy schema evolution, complete with benchmark comparisons and practical deployment guidance.
DaftData EngineeringLance
0 likes · 21 min read

This article examines how large‑language models can reconstruct the data‑warehouse DWD layer by automating ETL script generation, data cleaning, standardization, and cross‑table association, presenting three high‑frequency scenarios—structured data cleaning, multimodal data parsing, and intelligent table linking—along with tool selections, step‑by‑step procedures, real‑world case studies, and practical pitfalls.
AIDWDETL Automation
0 likes · 18 min read

Haier’s digital transformation leverages a cloud‑native, open‑source‑based multi‑modal data lake that unifies structured and unstructured industrial data, uses metadata models and knowledge graphs for governance, and provides AI‑ready services that balance performance, cost, and real‑time requirements.
AIMultimodal Databig-data
0 likes · 12 min read

The article recaps the Apache Hudi Asia Meetup hosted by JD, covering community updates, JD's data‑lake challenges, the upcoming Hudi 1.1 release, JD's architectural redesign, Kuaishou's real‑time lake adoption, and Huawei Cloud's deep optimizations, all aimed at building an AI‑native, real‑time lakehouse.
AI NativeApache HudiFlink
0 likes · 13 min read

This article series explores Tencent Cloud's Iceberg‑based batch‑stream integration, Apache Gravitino's unified metadata and lineage solution, Xiaohongshu's data‑architecture evolution for the Big AI Data era, and a practical Data+AI multimodal data‑lake implementation, highlighting challenges, architectural designs, and performance gains.
Multimodal Databig datadata lake
0 likes · 7 min read

This article collection examines Tencent Cloud’s Iceberg batch‑stream integration, AI‑driven game data governance, Apache Gravitino unified metadata and lineage, Xiaohongshu’s multimodal data‑lake evolution, and Volcano Engine’s Data+AI multimodal lake, highlighting architectures, techniques, performance gains, and practical implementations.
AI governanceGravitinoMultimodal Data
0 likes · 7 min read

This guide walks through constructing an AI‑powered multimodal data analysis platform on Hologres 4.0, covering PDF ingestion, vector and full‑text indexing, AI functions, stored procedures, and hybrid retrieval queries to enable enterprise‑level search and insight extraction.
AIHologresMultimodal Data
0 likes · 41 min read

Game‑TARS, a general‑purpose game AI from ByteDance's Seed team, replaces custom function calls with low‑level keyboard‑mouse actions, leverages massive multimodal data, sparse‑thinking and decaying‑loss algorithms, and achieves zero‑shot mastery across diverse games, surpassing top large models like GPT‑5 and Gemini‑2.5‑Pro.
AI TrainingMultimodal Datagame AI
0 likes · 10 min read

The article discusses the challenges of training ever‑larger AI models on internet‑scale data, critiques traditional batch ETL pipelines, and proposes a streaming data‑flow architecture with dynamic data selection and a shared‑memory/Alluxio middle layer to decouple data processing from model training, improving efficiency and scalability.
AI infrastructureMultimodal Datadata pipelines
0 likes · 20 min read

The talk reviews the rapid evolution of data technologies—from early database foundations and big‑data breakthroughs to the rise of generative AI—highlighting how Ant Group’s data platform is shifting from a cost‑efficiency focus to a value‑centric, multimodal, AI‑driven ecosystem.
Data EngineeringData PlatformsMultimodal Data
0 likes · 17 min read


21CTOSep 7, 2024 · Artificial Intelligence
The article explains how AI databases combine structured, unstructured, and vector data, integrate machine‑learning, NLP, and generative models, and why platforms like Vespa are emerging as open‑source solutions to meet the performance and scalability demands of modern generative AI applications.
AI DatabaseGenerative AIMultimodal Data
0 likes · 8 min read

This article reviews the rising importance of data quality for large models, explores data‑centric AI, large‑model pre‑training data engineering, and presents recent Fudan University research on using large models to improve data governance across multiple domains such as attribute normalization, geographic cleaning, compliance checking, and multimodal retrieval.
AIData EngineeringMultimodal Data
0 likes · 19 min read

This presentation details how Wing Payment leverages a large‑scale, multimodal knowledge graph and AI techniques—including computer vision, unsupervised and supervised learning, federated learning, and graph neural networks—to detect and mitigate fraud across payment, e‑commerce, and credit scenarios, while outlining system architecture, algorithmic approaches, case studies, and future research directions.
Financial ServicesGraph AlgorithmsMultimodal Data
0 likes · 17 min read


58 TechApr 16, 2021 · Artificial Intelligence
The article presents a comprehensive anti‑fraud framework that analyzes black‑market fraud characteristics, reviews conventional fraud‑mitigation methods, and proposes a multimodal graph‑neural‑network approach—leveraging device, behavior, and content similarity—to accurately identify fraudulent users on large‑scale internet platforms.
Graph Neural NetworksMultimodal Dataanti‑fraud
0 likes · 18 min read
