The article introduces xeasy-ml, a Python machine‑learning framework, explains its installation requirements, demonstrates how to initialize a project, run training, view result artifacts, and provides detailed instructions for online prediction using configuration files and code examples.
This article examines how startups can harness massive pre‑trained AI models such as GPT‑3, outlining the historical context, benefits of transfer learning, the steep costs and data‑alignment challenges, and strategic considerations when using cloud APIs versus self‑hosting.
The article outlines Baidu MEG's evolution of intelligent software testing through three AI-driven stages—computational, perceptual, and cognitive—showcasing how data and algorithms enhance test efficiency, risk perception, and automated decision‑making while also noting current limitations and future prospects.
This article presents a comprehensive overview of Huya Live's recommendation system, covering business background, system architecture, vector retrieval techniques, ranking pipeline, technical challenges, implementation details, and future outlook, highlighting scalability and performance optimizations.
This article details Alibaba's exploration of intelligent supply‑chain forecasting, covering scenario classification, three generations of prediction algorithms, the self‑developed Falcon model, performance evaluation, and real‑world cases such as Double 11 and live‑streaming, highlighting challenges and practical solutions.
The paper applies causal inference and tree‑based uplift modeling to identify coupon‑responsive riders, using T‑, S‑, and X‑Learners as well as a proprietary Treelift model that directly optimizes per‑user utility, achieving a 4.7% lift over manual rules and 2.3% over prior response models.
This article explains how Hello Mobility applies causal inference and tree‑based uplift models to improve the efficiency of hotel‑marketing subsidies, detailing background, problem formulation, various uplift learning methods, a customized TreeCausal (Treelift) algorithm, offline AUUC evaluation, online deployment, and future research directions.
This article introduces the Dxm Eros ultra‑large graph platform, explains its architecture, storage, analysis, modeling and visualization capabilities, and demonstrates how graph‑based machine learning is applied to fintech risk control, anti‑fraud, anti‑money‑laundering and automated audit workflows.
This article presents Beike's experience in applying intelligent advertising to the real‑estate industry, detailing business background, key technical challenges, multi‑task conversion modeling, GEO user targeting, delayed‑feedback calibration, and ROI‑driven budget allocation to improve ad effectiveness.
This article explores the classification of digital humans, examines how they transform user experience, and presents a detailed design practice for AI‑driven virtual bank tellers, covering visual identity, interaction flow, ergonomic screen layout, and future considerations for natural multimodal interfaces.
This article details JD's heuristic question‑answering framework for intelligent customer service, covering its pre‑consultation prediction, in‑consultation associative input, post‑consultation recommendation modules, underlying algorithms, deployment results, and future enhancement directions.
The article details Meituan's five‑year journey in instant logistics, describing the evolution from early vertical services to micro‑service and cloud‑native architectures, the technical challenges of massive order‑rider matching, high availability, AI‑driven optimization, and the operational practices adopted to ensure scalability and reliability.
The paper presents RISA, a reference‑guided image synthesis assessment model that learns to score the quality of a single generated image without human‑labeled data by leveraging GAN intermediate outputs, pixel‑wise interpolation, multiple binary classifiers, and contrastive learning, achieving results comparable to human perception and earning an AAAI 2022 oral presentation.
The article reviews three notable 2022 machine‑learning advancements—CatBoost, Amazon SageMaker's DeepAR forecasting, and the low‑code PyCaret library—explaining their unique advantages, typical use‑cases, and why they are expected to gain traction among data scientists.
The article details how DaJia's recommendation system progressed through three architectural versions—V1.0's simple strategy‑factory design, V2.0's vertical business split and configurable pipeline, and V3.0's dynamic configuration service and modular pipeline—addressing scalability, fault isolation, and personalization challenges while outlining future directions for explainable AI recommendations.
This report details how the Search Recommendation team at 58.com upgraded their deep learning ranking model for recruitment by adding multi-valued and semantic vector features, integrating conversion sequences, employing feature‑crossing techniques, optimizing offline data pipelines, and planning future multi‑scene improvements to boost CTR and relevance.
The article analyzes Gartner's 2026 metaverse forecast, outlines three development stages, examines the driving technologies such as 5G, AI, blockchain and VR/AR, and breaks down the industry chain into underlying tech, interaction devices, development platforms, and applications to show how the metaverse will transform work, consumer experiences and the broader internet ecosystem.
The article explains Gartner’s augmented analytics paradigm—automating data preparation, insight discovery, and sharing via AI/ML—and details Alibaba Mama’s AutoInsight system, whose four‑layer architecture delivers automated anomaly detection, metric contribution analysis, personalized recommendations, and real‑time warnings, illustrated through advertising and mobile placement use cases, highlighting its potential to boost marketing efficiency and insight automation.
This article reviews the motivation, methods, and results of graph-based self‑supervised pretraining for molecular data, introduces the ChemRL‑GEM model that incorporates 3‑D structural information, and demonstrates its superior performance on ADMET, affinity prediction, and benchmark competitions using the PaddleHelix platform.
This article introduces two AI-driven anime portrait tools—Magical Anime Portraits and the open‑source SofGAN project—explaining their workflow, underlying GAN technology, related research, and how they enable users to create unique, customizable anime avatars without manual drawing.
This article reviews recent research on e‑commerce search, covering transformer‑based complementary rankings, Alibaba's cognitive concept net and its extension, joint deep retrieval with product quantization, personalized semantic retrieval, multi‑granularity deep semantic retrieval, and graph‑attention networks for long‑tail shop search.
This article summarizes a cloud‑native expert’s talk, covering the definition and community origins of cloud native, CNCF’s role, current industry trends such as distributed cloud, AI workloads, FinOps, core technologies like containers, microservices, serverless, and Alibaba Cloud’s product ecosystem and open‑source contributions.
The article explores the evolution of the Internet from IPv4 to IPv6 and IPv6+, highlighting how next‑generation protocols, AI, SDN, and enhanced security create a more convenient, open, intelligent, and trustworthy digital ecosystem.
The article showcases three AI‑driven sports video solutions—real‑time figure‑skating action recognition with ST‑GCN, multimodal video classification merging text, image and audio via ERNIE and TextCNN, and automated football highlight clipping using TSN‑BMN‑LSTM—each achieving over 85% accuracy, fully open‑source on PaddlePaddle with one‑click notebooks and a live developer session.
The article examines post‑COVID‑19 remote work trends, highlighting the growing importance of cloud migration, DevOps, AI‑assisted development, low‑code platforms, and practical management techniques to boost productivity and employee well‑being in a permanently distributed workforce.
On February 14, AMD completed its landmark $35 billion all‑stock purchase of Xilinx, a move that expands AMD’s product portfolio with FPGA and adaptive SoC technologies, strengthens its position in AI, data‑center and edge markets, and signals a strategic push to compete with Intel.
This article details the evolution of an e‑commerce recommendation system through three architectural versions, highlighting the initial simple design, the challenges that prompted vertical and horizontal splits, the introduction of a configurable pipeline and AB testing, and the final micro‑service‑based, dynamically configurable V3.0 architecture.
Baidu’s AI‑driven sign‑language digital human combines ultra‑accurate speech recognition, specialized translation, and precise gesture‑generation models—backed by extensive motion‑capture data and expert validation—to deliver 24‑hour, high‑fidelity signing for millions of hearing‑impaired users, showcasing inclusive AI communication.
This step‑by‑step guide shows how to use Huawei Cloud ModelArts and the U‑2‑Net AI model to create a personalized portrait for Valentine’s Day, covering account registration, environment configuration, data and code download, dependency installation, generation, and result display.
Researchers propose a causal learning framework IV4Rec to separate causal and non-causal factors in recommendations using search data.
The article presents a case study of how Alibaba Cloud’s network operations team tackled the massive, unpredictable traffic spikes of the 2021 Double 11 shopping festival by identifying transaction‑promotion traffic risks early through AI‑powered analysis, overcoming the limitations of manual rule‑based detection, and achieving precise, automated capacity risk control.
This talk by senior algorithm engineer Su Yonghao from 58.com explains how embedding vectors are applied to commercial search and recommendation scenarios, covering traffic contexts, mainstream embedding algorithms, practical implementations, query expansion, relevance modeling, and system architecture, and shares evaluation results and Q&A.
This article provides a comprehensive, data‑driven overview of front‑end technology trends from 2021 to 2022, covering programming language popularity, framework adoption, build tools, low‑code platforms, AI‑assisted development, cross‑platform solutions, serverless architecture, DevOps, 5G impacts, WebAssembly, and future directions for developers.
The Alibaba Mama technical team secured five paper acceptances at The Web Conference 2023, presenting advances in unbiased delayed‑feedback conversion modeling, uncertainty‑regularized knowledge‑distilled CVR debiasing, feature‑aware probability calibration, coordinated two‑stage ad auctions, and scalable decoupled graph neural networks for large‑scale e‑commerce retrieval.
This article explains why API design is crucial for AI open-source software, outlines new challenges, presents basic design principles, and offers practical guidelines, examples, and best practices to create user-friendly, stable, and extensible interfaces that lower learning costs and foster ecosystem growth.
This article describes the progressive redesign of an e‑commerce recommendation platform—from a simple strategy‑factory V1.0 through a vertically split V2.0 to a fully configurable, pipeline‑driven V3.0—highlighting architectural challenges, Redis clustering, dynamic configuration, recall and prediction services, and future directions for fine‑grained, explainable recommendations.
This article introduces the Dxm Eros ultra‑large graph platform, detailing its architecture, storage, analysis, modeling, and visualization modules, and demonstrates how graph machine‑learning techniques are applied to financial risk control, fraud detection, anti‑money‑laundering, and automated credit review.
This presentation explains the fundamentals of financial risk control, outlines the data and AI technologies needed across the pre‑loan, in‑loan, and post‑loan stages, and showcases visual‑analytics and federated‑learning case studies that improve model interpretability, monitoring, and enterprise risk management.
This article introduces causal analysis, outlines its major challenges such as correlation versus causation, confounding factors, and selection bias, explains a three‑step framework (association, intervention, counterfactual), and details how Beike applied these principles in a smart client‑management tool with rigorous A/B experiments.
Arm's newly revealed Armv9 architecture introduces major ISA extensions, a confidential compute model, AI‑focused vector capabilities, and performance road‑maps that together aim to unify mobile, server and future workloads while strengthening security and software ecosystem compatibility.
This article compares PyTorch and TensorFlow in 2022 across model availability, deployment ease, and ecosystem support, using data from HuggingFace, research papers, and industry tools, and offers tailored recommendations for industry engineers, researchers, educators, career changers, hobbyists, and beginners.
This curated list showcases Alibaba Mobile Technology's standout 2021 articles, covering HarmonyOS rendering engine analysis, Youku's HarmonyOS development series, Gaode Map performance optimization, AndroidX migration, Flutter performance tricks, security production systems, QUIC protocol adoption, and cutting‑edge OCR and AI advancements.
InfoWorld's 2021 BOSSIE Awards highlight 29 standout open‑source projects—from front‑end frameworks like Svelte to cloud‑native tools such as Minikube, AI platforms like Hugging Face, data‑engineered solutions including Presto and Apache Arrow, and many more—offering developers a curated snapshot of the most influential software of the year.
This article explains how Kuaishou's Y‑Tech leverages human detection, keypoint localization, and image‑deformation algorithms such as stretching, triangulation and liquify, together with background‑distortion correction, to deliver seven stable, natural body‑shaping effects for short‑video applications.
Amid post‑pandemic uncertainty, Baidu Research outlines 2022 AI breakthroughs—from large‑model advancements and cross‑modal knowledge enhancement to AI‑driven scientific discovery, privacy computing, quantum hardware, autonomous driving, green AI, and inclusive technologies—highlighting how these trends will reshape industries and society.
This article presents Meituan's search ranking system, detailing its multi‑business sorting architecture, layered ranking pipeline, quota and fine‑ranking models, aggregation modeling techniques, and supporting platforms such as Lego and Poker, while also sharing practical insights and recruitment information.
This talk presents Tencent's multimodal content understanding framework for search, covering hierarchical content features, large‑scale ranking, fine‑grained image semantic vectors, video and document analysis, quality detection, duplicate removal, and future directions in AI‑driven search.
This article explains Kuaishou’s Y‑tech body‑beautification pipeline, detailing how proprietary human pose detection, key‑point localization, and image‑warping techniques such as stretching, triangulation, and liquify are combined to create stable, natural effects like long‑leg, slim‑waist, and swan‑neck, while minimizing background distortion.
The article introduces Netease Thunderfire UX's big data team and its multidisciplinary approach to game user experience, covering topics such as mathematics, engineering, product practice, AI, visualization, platform development, data analysis, and tool creation for game data professionals.
The article analyzes how hyper‑converged data‑center networking unifies compute, storage, and HPC on a lossless Ethernet fabric, addresses AI‑era performance bottlenecks, compares RDMA over Ethernet with InfiniBand, and outlines the core metrics, value, and key technologies that enable zero‑loss, low‑latency, high‑throughput operation.
This article presents a comprehensive technical overview of Alluxio, covering its role as a distributed data orchestration layer for AI workloads, core features such as caching and unified namespace, performance challenges in large‑scale machine‑learning pipelines, and the extensive optimizations and testing performed at Tencent to achieve high throughput and scalability.
This article explains how KubeDL, Alibaba’s open-source Kubernetes-based AI workload framework, extends standard container networking with HostNetwork support to eliminate overlay overhead, detailing the benefits, challenges, configuration steps, and performance gains for large-scale distributed training.
This article explains video frame interpolation, compares algorithm categories, and deep‑dives into Kuaishou’s KFRUC method—covering its hybrid motion estimation, occlusion detection, and motion compensation techniques that enable smooth slow‑motion playback without stutter.
This article provides a comprehensive overview of fraud detection using graph neural networks, covering background definitions, fraud categories, GNN application steps, a timeline of key research papers, practical challenges, solutions, and a collection of open‑source resources and datasets.
This case study details 58.com’s 11.0 product overhaul, covering six months of research, data analysis, live‑stream and short‑video integration, AI/AR/VR applications, redesign of the homepage layout, and the evolution of the LOCO Design System to improve user efficiency and service reach.
The team built a lightweight AI‑FAAS platform that migrates C++ ad‑algorithm services to a Java‑centric, cloud‑native environment, encapsulates core operators as DolphinSQL plugins, and defines real‑time feature pipelines via standardized SQL, cutting service creation from weeks to days, enabling rapid iteration, high operator reuse, and near‑instant feature delivery.
The article outlines a comprehensive framework for evaluating enterprise digital maturity, describes the seven key characteristics of a digital enterprise—including customer‑centric organization, diversified capabilities, AI‑driven insight, agile practices, cloud‑5G integration, and evolving IT roles—and offers practical solutions to common transformation challenges.
This tutorial explains how to create a Siri‑style conversational robot in Python by configuring the environment, describing the speech‑recognition and chatbot principles, and showing the implementation that uses Baidu speech recognition and the Turing chatbot API.
This article details the 58.com WPAI machine learning platform's architecture and the optimizations applied to training task scheduling, inference service elastic scaling, and offline‑online resource mixing, demonstrating how these techniques significantly improve GPU/CPU utilization and inference performance across both GPU and CPU environments.
The article outlines five major trends—neural-network-enhanced search, semantic search, document understanding, image and voice search, and knowledge graphs—that are transforming enterprise use of unstructured data by leveraging AI to deliver precise, context-aware answers and insights.
This article reviews the fundamentals of fraud, surveys the evolution of graph neural network research for fraud detection, outlines practical application steps, discusses key challenges such as disguise, scalability, and label scarcity, and provides representative papers, new research directions, industrial case studies, and open-source resources.
Alibaba DAMO Academy’s 2022 report identifies ten emerging technology trends—from AI‑driven scientific research and large‑model co‑evolution to silicon photonic chips, green‑energy AI, flexible perception robots, high‑precision medical navigation, universal privacy computing, star‑ground computing, cloud‑network‑edge fusion, and XR internet—detailing their impact and expected development over the next few years.
The article provides a comprehensive overview of the ARM ecosystem in 2021, covering its business model, key partnerships, automotive and AI applications, new Cortex and Neoverse processors, Chinese ARM core families, cloud deployments, and the strategic moves of major technology companies.
On Jan 7, 2024, BIT’s “Hundred Lectures” will feature Assistant Professor Hanwang Zhang presenting his group‑theoretic self‑supervised representation learning work, including the IP‑IRM method that iteratively partitions data and applies invariant risk minimization to achieve fully disentangled visual features, with the session streamed via Tencent Meeting.
This article outlines the major DevOps trends—including micro‑service architecture, serverless computing, low‑code platforms, Kubernetes evolution, DevSecOps, and AI integration—that are expected to drive faster delivery, higher quality, and greater agility in software development and operations.
The article outlines the three major visions for future networks—Universal Connectivity, Compute‑Network Convergence, and Precise Networking—detailing their technical requirements, security and mobility challenges, and the design principles and architectural models needed to achieve them by 2030.
The article examines Parag Agrawal’s transition from Twitter CTO to CEO, explores the growing trend of CTOs moving into top executive roles, and offers practical advice for technical leaders aiming to broaden their business and financial expertise.
The article explains what the Metaverse is, why it has become a hot concept, outlines its six‑layer technological architecture, and discusses the current state and future opportunities of the Metaverse in China, including key players and emerging sub‑domains such as VR, AR, gaming, and virtual humans.
This article explains how to write automated test case scripts for game development, distinguishing good and bad test cases, detailing script language selection, state‑machine based Lua examples, timer usage, AI‑driven testing, performance and version checks, and best practices for maintainable, reliable automation.
This article reviews Baidu's knowledge‑enhanced semantic understanding models, detailing the evolution from early semantic techniques to ERNIE 1.0, 2.0 and the large‑scale ERNIE 3.0, its architecture, training strategies, performance benchmarks, and real‑world applications across industry.
This article reviews the challenges of extracting structured table data from images, compares two‑stage and end‑to‑end OCR approaches, evaluates four state‑of‑the‑art table‑recognition models (SPLERGE, CascadeTabNet, TableMASTER, UnetTable), and presents a practical deployment workflow with performance metrics.
This article explains the concept of feature engineering, illustrates it with a height‑weight classification example, compares kernel‑enhanced models to handcrafted features like BMI, discusses its impact on model performance, and highlights practical tips and domain‑specific considerations.
OpenCV 4.5.5 introduces audio support in VideoCapture, updates SOVERSION handling, adds OpenVINO 2021.4.2 LTS compatibility, expands ONNX test coverage, upgrades protobuf, optimizes for RISC‑V, and enhances the G‑API module with numerous vectorized kernels, SIMD scheduling, and various bug fixes.
This article reviews the ten most notable Python libraries for machine learning in 2021, covering tools for handling nested data, notebook versioning, lightweight model demos, data management, multimodal augmentation, model monitoring, object detection, fast inference, time‑series forecasting, and neural search.
An in‑depth 2022 comparison of PyTorch and TensorFlow evaluates model availability, deployment ease, and ecosystem support, showing PyTorch dominates research while TensorFlow excels in deployment, and offers tailored recommendations for industry professionals, researchers, educators, career changers, hobbyists, and beginners.
Zhejiang Mobile’s fault‑center AIOps system passed the first official assessment by China’s CAICT, showcasing advanced AI‑driven IT operations, detailing the evaluation process, interview insights, and the broader AIOps maturity model shaping future intelligent operations.
This talk reviews recent progress in neural‑symbolic learning and multimodal knowledge discovery, highlighting examples such as GPT‑3 reasoning failures, the need for symbolic knowledge, historical developments, various integration methods, challenges in multimodal knowledge graphs, and future research directions.
This article reviews three consecutive works from Alibaba DAMO Academy on compressing and distilling large pretrained language models—AdaBERT, L2A, and Meta‑KD—detailing their motivations, neural‑architecture‑search‑based designs, loss formulations, experimental results, and insights from a Q&A session.
This article explains how Tencent KanDian analyzes user behavior by introducing the product, describing common clustering scenarios, reviewing traditional unsupervised methods, and detailing advanced path‑based approaches such as N‑gram and action2vec, while discussing their advantages, limitations, and practical applications.
KeSpeech, a benchmark jointly released by Beike AI and Tsinghua University at NeurIPS 2021, provides a massive Chinese Mandarin dialect dataset covering 30,000 speakers from 34 cities, supporting speech recognition, speaker verification, dialect identification, and voice conversion tasks, and includes rich multi‑scenario and parallel corpora for advanced research.
This case study details the design of the Digital Palace Museum WeChat mini‑program, covering its cultural goals, simplified visual language, clear content categorisation, immersive and warm UI for all ages, AI‑driven voice assistance, AR navigation, and real‑time crowd‑flow features.
This talk by Huawei senior engineer Xu Chi explores the challenges of drug screening, AI-driven drug discovery practices, and how personalized federated learning combined with Huawei Cloud's high‑performance computing accelerates pharmaceutical research, including case studies, platform services, and collaborative efforts.
The talk by Huawei senior engineer Xu Chi explores the challenges of drug screening, recent AI-driven advances in pharmaceutical research, and how personalized federated learning on Huawei Cloud accelerates drug discovery, including case studies, platform architecture, and large-scale virtual screening services.
At the WISE2021 China Digital Innovation Summit, Zhongtong Express CTO Zhu Jingxi detailed the company's digital transformation journey, highlighting the impact of electronic waybills, data-driven operations, AI routing, and privacy security on reshaping the logistics supply chain and boosting efficiency.
This article presents OPPO's XiaoBu Assistant platform, detailing its low‑code workflow for business domain modeling, multi‑type NLU (model‑based, retrieval‑based, and QA), componentized core services, large‑scale text processing, vector retrieval, and flexible dialogue management that together enable a complete skill lifecycle from development to online optimization.
This article defines virtual characters, outlines their market growth and industry chain, showcases leading products and solutions, and details the technical research—including AI-driven animation, rendering pipelines, scene orchestration, and big‑data algorithms—being pursued by Alibaba's front‑end team.
The article outlines the three major visions for future networks—global interconnection, compute‑network convergence, and precision networking—detailing their technical requirements, intrinsic security, mobility management, and design principles, and proposes a layered architecture that extends network capabilities to both ends of applications.
This article presents an overview of speech translation, discusses its motivations and applications at ByteDance, compares cascade and end‑to‑end modeling approaches, introduces advanced encoder and decoder designs such as LUT, Chimera, and COSTT, outlines progressive multi‑task training and data‑augmentation strategies, and shares experimental results and Q&A.
This article presents five concise, widely applicable techniques—adversarial training with FGM, exponential moving average (EMA), test‑time augmentation (TTA), pseudo‑label learning, and special‑sample handling via nearest‑neighbor retrieval—to reliably improve model performance with minimal code changes.
Zhou Jiaxin explains that as IoT moves from universal connectivity to intelligent integration, Tencent Cloud IoT’s “Lianlian” platform tackles cost, efficiency and ecosystem gaps by embedding content, AI, big‑data and WeChat services into eight modular solutions, enabling rapid, cross‑industry smart applications.
This article describes an AI-driven portrait enhancement system that restores degraded facial images by simulating degradation, constructing paired datasets, and employing a StyleGAN‑based generator combined with face‑parsing masks, detailing the pipeline, model architecture, training losses, and achieved high‑quality results.
The article explains how hyper‑converged data center networking, driven by AI‑era demands for zero‑packet loss, ultra‑low latency, and high throughput, replaces traditional separate networks with a unified Ethernet‑based solution that simplifies operations, reduces cost, and enhances performance.
This talk presents the design and implementation of 58’s commercial recruitment recommendation system, covering the characteristics of the app’s recommendation scenario, system architecture, region‑based and behavior‑based recall methods, and coarse‑ and fine‑ranking models with various optimizations and future directions.
On December 6, Kuaishou and Tsinghua University’s joint team secured first place in the NeurIPS'21 Billion‑Scale Approximate Nearest Neighbor Search Challenge by leveraging a FAISS‑optimized, memory‑efficient KST_ANN algorithm that achieved over 6% higher recall on multiple billion‑scale datasets, showcasing the practical impact of large‑scale vector retrieval in AI‑driven services.
Image clarity impacts daily life, from personal memories to security, and Kuaishou’s Y‑Tech team tackles degradation by constructing paired low‑high quality datasets and a style‑based AI model that leverages facial masks to restore high‑definition portraits, preserving identity while enhancing detail.
This guide outlines six Python advancement directions—full‑stack development, web scraping, data analysis & mining, AI & machine learning, DevOps automation, and testing—detailing essential skills, typical salary ranges, and how each role fits the modern job market.
The article examines the concept of "new infrastructure" in China, outlines the evolution of the Internet toward a third generation, discusses candidate future network architectures such as SDN, CCN, and XIA, and reviews emerging data‑center networking technologies like large‑scale L2, VXLAN, virtual switching, and large‑scale switching, highlighting their role in supporting AI, big data, and cloud computing workloads.
This talk presents DAMO Academy’s end‑to‑end NLG pipeline for e‑commerce, covering the XGeneration content‑reproduction system, controllable short‑text generation models (PKM, PSCN), KG‑to‑text conversion, metaphor generation, and new‑media extensions such as short‑video script creation and intelligent video editing, together with experimental results and practical applications.
Paddle.js OCR SDK brings Baidu’s lightweight PaddleOCR models to web browsers, offering init() and recognize() APIs that load the ch_PP-OCRv2 detection (DB) and recognition (CRNN with bidirectional LSTM) models in parallel, achieving 258 ms detection, 60 ms recognition, 0.52 F‑score, and a combined size under 12 MB.
Alibaba’s DAMO Academy and Tsinghua University introduced M6‑UFC, a non‑autoregressive multimodal transformer that unifies arbitrary text and image controls to generate high‑quality, editable fashion designs, dramatically reducing carbon emissions and outperforming GAN‑based models in fidelity and relevance while accelerating production speed.
