The 2017 Transformer model introduced by Vaswani et al. revolutionized machine translation by relying solely on attention mechanisms, outperforming traditional RNN and CNN approaches through parallel processing and improved contextual understanding.
The article provides a comprehensive overview of RISC‑V, covering its origins amid trade‑war pressures, open‑source technical characteristics, growing global and Chinese ecosystem, comparisons with ARM and x86, and its emerging role in low‑power, customizable chips for IoT and AI applications.
This article describes how 58.com’s AI Lab integrated machine‑learning and recommendation techniques into its CRM system, redesigning sales workflows, introducing the “Michigan” model, and deploying XGBoost and MMoE models to boost key metrics such as transfer rate and 60‑second effective call rate, achieving significant performance gains.
This article curates ten standout presentations from the 2020 NVIDIA GTC China conference, highlighting how DPU technology accelerates finance, 5G, edge computing, data science and high‑performance networking across industry‑leading experts.
In 2021 big data will keep expanding, driven by five key trends—cloud migration, augmented analytics, IoT integration, Data‑as‑a‑Service growth, and the rise of edge computing—that together help developers make data‑driven decisions and innovate across industries.
This article explains how customer‑service chatbots use well‑structured knowledge bases and advanced sentence‑similarity algorithms to achieve over 90% question‑matching rates, covering knowledge‑base construction, maintenance, and the core AI models that power accurate responses.
In a 2020 talk, Kevin Kelly explores how the rapid rise of smartphones and emerging technologies like smart glasses and mixed reality will reshape learning, emphasizing the need to master "Techno Learning," collaborate with AI, and focus on asking better questions for the next 5,000 days.
The article introduces three major instruction‑set families—CISC (x86), RISC (ARM) and the ultra‑lightweight RISC‑V—explaining how RISC‑V, as the fifth‑generation RISC architecture, balances data throughput and speed, making it an ideal, self‑controlled solution for China’s AI and IoT development despite its still‑growing ecosystem.
This week’s developer newsletter highlights China’s 11.3‑billion‑parameter Wenhui AI model, Rokid’s dual‑camera MR glasses, the TypeScript 4.2 beta release, a Go generics proposal for 1.18, Docker’s 2021 priorities, the Industrial Internet action plan, a robust DARTS paper from ICLR 2021, and the open‑source TextBox library for text generation.
This article presents the 2020 Python library ranking, explaining the selection criteria and highlighting ten standout libraries—ranging from CLI tools like Typer and Rich to AI‑focused frameworks such as Hydra, PyTorch Lightning, Hummingbird, and HiPlot—plus several honorable mentions.
The article examines how iQIYI leverages AI across the video ecosystem—from intelligent material search, old‑film restoration, and voice cloning to virtual idols, XR production, and AI‑driven advertising—to boost creator efficiency, enhance user experience, and accelerate industry-wide digital transformation.
In a comprehensive interview, Elon Musk shares his views on artificial intelligence, the future of autonomous vehicles, battery resource challenges, the strategic importance of Tesla's Berlin factory, his personal wealth philosophy, the need for multi‑planetary humanity, and his definition of himself as an engineer.
In this interview, Tencent Classroom’s front‑end lead Jay shares how his team built a systematic front‑end engineering efficiency framework, covering the evolution from basic build tools to full‑stack CI/CD, AI‑driven automation, and practical advice for new developers.
The article explains how EpiK uses blockchain‑based storage, incentive mechanisms, and DAO governance to build a decentralized, open‑source knowledge‑graph platform that addresses AI data needs, coordination challenges, and encourages participation from experts, miners, and bounty hunters.
Alibaba’s DAMO Academy unveils its 2021 top ten technology trends, highlighting breakthroughs in third‑generation semiconductors, post‑quantum computing, flexible carbon‑based electronics, AI‑driven drug discovery, brain‑computer interfaces, autonomous data processing, cloud‑native architectures, smart agriculture, industrial IoT, and intelligent city operation centers.
The article reviews 2021 front‑end predictions, highlighting a shift toward stability, deep‑water exploration of serverless platforms, intelligent UI, HTTP‑import modules, and AI‑assisted development, while showcasing several open‑source projects that illustrate these emerging directions.
This tutorial shows how to create a civilian‑grade AI for the mobile game Honor of Kings by mirroring the phone with scrcpy, generating action commands from game screenshots using a Transformer model, and executing them via minitouch on Android devices.
This article describes how NetEase built a real‑time log intelligent classification system using Flink and AI algorithms, detailing the challenges of massive log volumes, the Drain template‑extraction method, algorithm workflow, performance results, and a practical case study that demonstrates reduced alert storms and faster issue diagnosis.
Zhang Zhengyou, head of Tencent AI Lab and Robotics X, has become the company’s inaugural T17‑level researcher, a distinction earned through pioneering work in AI‑driven agriculture, virtual humans, facial recognition, speech front‑ends, advanced robotics like the Jamoca quadruped, and contributions such as the Go AI “Absolute Art”.
This article examines the technical evolution of POI expiration detection, covering attribute‑based, behavior‑based, and human‑place relationship mining methods, their machine‑learning models, and how they collectively improve map freshness and user experience at scale.
From the rise and fall of Flash to modern front‑end video playback techniques, performance optimization strategies, AI recommendation fundamentals, CLI design best practices, and a glimpse into game development and algorithm analysis, this article surveys diverse cutting‑edge technologies shaping today’s software landscape.
The Chinese "Grain Shield" project, led by Beijing University of Posts and Telecommunications, uses AI, deep‑learning and Huawei Cloud services to automatically detect pests and mold in grain silos, dramatically cutting post‑harvest loss, reducing chemical use, and demonstrating a scalable smart‑agriculture solution across dozens of provinces.
This article reviews four mainstream model compression and acceleration methods—structural optimization, pruning, quantization, and knowledge distillation—explaining their principles, implementations, and performance, and presents practical examples such as DistillBERT, TinyBERT, and FastBERT with comparative results.
The article examines how electrification, intelligence, connectivity and sharing drive new autonomous‑driving requirements, detailing challenges in vehicle platform architecture, sensor technologies (camera, lidar, radar), high‑performance computing, integration, and rigorous test validation needed to achieve pre‑installed mass‑production autonomous vehicles.
This article explains how developers can quickly create and run graph algorithm tasks, such as GraphSage, on Tencent Cloud's TI-ONE platform by leveraging its drag‑and‑drop workflow, built‑in Angel algorithms, and flexible resource management to accelerate AI model deployment.
This article explores the challenges of enterprise knowledge management and presents vertical‑domain knowledge graph construction techniques based on OpenIE, covering data analysis, extraction methods, SPO modeling, both closed‑ and open‑domain approaches, and practical applications such as chatbots, search and intelligent QA.
AI-driven super-resolution, once limited to academic research, is now tackling real-time video communication challenges by evolving from early interpolation methods to deep learning models, addressing issues of model size, generalization, and real-world degradation, while lightweight networks and encoding-aware techniques promise practical deployment.
This week’s developer newsletter covers the release of a unified 3D data file specification, Adobe Flash’s end of life, deepin OS 20.1 updates, China Telecom’s quantum‑secure call service, a call for new AI benchmarks, Windows 10 Cloud PC details, YOLObile’s 19 FPS mobile detection breakthrough, and the RealFormer transformer improvement.
This weekly roundup covers a Chinese open‑source JD seckill script topping GitHub trends, a severe Btrfs slowdown in Linux 5.10, native Linux WeChat for UOS, Flash Player removal, Mi 11 kernel open‑source, Ruby 3.0 launch, record app‑store spending, Linux ported to Nintendo N64, DeepMind's MuZero AI, and Huawei's leading Linux‑5.10 contributions.
This talk explains Alibaba’s P2C (PRD‑to‑Code) system, detailing how extending design‑to‑code pipelines to product requirement documents, combined with intelligent labeling and logic‑point extraction, enables front‑end engineers and product designers to automate code generation and accelerate development across the entire product lifecycle.
Tencent AI Lab’s WeKick AI clinched the inaugural Google Football Kaggle championship with a score of 1785.8, outpacing over 1,100 research teams by leveraging a migrated full‑stack architecture, custom reinforcement‑learning frameworks, GAIL imitation learning, and a multi‑style League training pipeline that together showcase the power and generality of deep reinforcement learning for complex multi‑agent tasks.
In early 2020, a small team of Huazhong University students partnered with Huawei to build an AI‑driven quantitative medical assistance system that processed thousands of CT scans in days, delivering rapid COVID‑19 diagnosis, severity assessment, and progression prediction for hospitals worldwide.
This talk explains Alibaba’s P2C product that extends design‑to‑code automation to PRD‑to‑code, detailing its layered architecture, annotation workflow, logical‑point extraction, code generation, and roadmap toward AI‑driven zero‑annotation development for front‑end engineers and product designers.
The article traces L4 autonomous-driving perception from early rule-based point-cloud methods through data-driven deep-learning models to emerging self-learning, multi-task systems, and highlights four key hurdles—model generalization and explainability, robust multi-sensor fusion, real-time compute limits, and proper uncertainty handling—calling for integrated AI, engineering, and data solutions.
This article examines the unique properties of novel literature and the difficulties of tag‑based recommendation, then details multi‑modal feature representation, dual‑tower semantic modeling, clustering, and YouTube‑style DNN recall techniques used to improve free novel recommendation systems.
This article introduces the typical applications of intelligent question answering, compares chat‑type, knowledge‑type and task‑type bots, and then details the end‑to‑end architecture, knowledge‑base construction, semantic‑equivalence modeling with BERT‑BIMPM, knowledge‑graph QA pipelines, and NL2SQL techniques, concluding with practical deployment insights.
This talk outlines 58.com’s risk control evolution, detailing the platform’s four development stages, the challenges of fraud, fake traffic, and content abuse, and how architecture, algorithms, and operational strategies have been refined to achieve high‑throughput, intelligent auditing.
The article outlines DiDi’s six‑year autonomous‑driving research, describing the three‑layer hardware‑onboard‑cloud infrastructure, key trade‑offs such as rapid iteration versus functional safety, sensor resolution versus compute, and hardware performance versus automotive‑grade reliability, and presents a staged evolution roadmap toward fully safe, driverless operation.
On January 9, 2021, Meituan hosted an unmanned‑delivery technical salon in Beijing where experts presented cutting‑edge AI research—including the CenterMask instance‑segmentation method, 3D geometry‑aware camera localization, multi‑agent trajectory prediction with attention‑based spatio‑temporal graphs, real‑time stereo visual‑inertial odometry calibration, and self‑supervised depth‑pose learning for dynamic scenes.
This article explains how Tencent's Angel platform powers game recommendation on Steam, Wegame, and the Tesla platform, comparing traditional linear tag‑based and collaborative‑filtering methods with the non‑linear DeepFM model and offering practical deployment tips.
The presentation outlines the AI engineering efficiency platform covering algorithm metric and evaluation, micro‑service performance testing, and dataset management architectures, detailing business pain points, platform‑wide improvements, technical designs, real‑world demos, and future directions to achieve accurate, fast, and stable AI services.
The article explains how JD Cloud’s federated learning platform enables secure, privacy‑preserving collaborative AI across isolated data sources by using encrypted distributed training, flexible model architectures, and a range of algorithms, while highlighting its architecture, security mechanisms, deployment speed, and real‑world industry successes.
The article reviews the rise of AI‑driven intelligent writing robots, outlines industry use cases, explains two main generation methods, and details how 58.com leverages massive structured data to automatically produce narrative, emotion‑rich listings for vehicles, housing, and jobs, enhancing user engagement.
tinygrad, an open‑source autograd tensor library by George Hotz, offers a compact PyTorch‑like experience in fewer than 1000 lines, with easy installation, GPU support via PyOpenCL, full EfficientNet inference, and extensible optimizers for rapid neural‑network prototyping.
The article explains how the open‑source Depix tool uses AI to decode pixelated or censored text, outlines its algorithm based on De Bruijn sequences, provides usage instructions, and also highlights the related PULSE AI up‑sampling technique for restoring low‑resolution images.
This week’s tech roundup covers Google’s two‑day service outage, the removal of over 60% of Pornhub videos, the launch of Rocky Linux, Linus Torvalds’ kernel submission deadline, Docker Desktop 3.0 updates, AI‑controlled U‑2 flight, major antitrust fines and the release of HarmonyOS 2.0 Beta, among other industry highlights.
The article analyses how the COVID‑19 pandemic accelerated global e‑commerce penetration, compares the evolution of cross‑border platforms, and outlines future opportunities in user acquisition, product selection, inventory forecasting, pricing and branding, emphasizing data‑driven and AI‑enabled operations to sustain growth.
This article explains how voice assistants rely on NLP's three processing layers, examines the shortcomings of the traditional DIS semantic structure, introduces the hierarchical CMRL schema with its six element types, and presents two neural models—copy‑write seq2seq and seq2tree—for accurate semantic parsing of complex commands.
The paper presents the Correlation Matching Feature Transfer (CMFT) network, which automatically colors line‑art animation frames by extracting line‑art features, computing pixel‑wise similarity to a colored reference, and transferring color features, achieving superior PSNR/SSIM scores and ~0.7 s per frame while cutting manual labor by roughly 30 %.
This article outlines SuNing's unmanned store technology, comparing its data-driven, product selection, and customer experience advantages over traditional shops, and detailing AI-powered applications such as face detection, target tracking, image recognition, and 3D reconstruction that enable 24‑hour service, intelligent merchandising, and precise customer analytics.
This article explains how KAIFX, a synthetic data platform built on computer graphics and AI techniques, tackles challenges of data scarcity, realism, labeling bias, and management to boost AR and 3D face reconstruction model performance.
This article explains the scale of online ad fraud, outlines common advertising billing models, describes how fake traffic generates revenue, defines invalid clicks, and presents a comprehensive anti‑fraud system that combines rule‑based methods, feature engineering, and AI models such as TextCNN, BiLSTM, BERT, Wide&Deep and GraphSage to identify and block fraudulent ad clicks.
This article recounts the JD.com 11.11 tech salon presentation where architect Shi Weihang explains how the NeuHub gateway was redesigned to handle over two billion TTS service calls, detailing the challenges, scaling techniques, and architectural decisions that enabled such massive traffic.
During JD.com's 11.11 shopping festival, engineers tackled billions of transactions by evolving a five‑year architecture roadmap, implementing robust messaging and delay queues, deploying smart NICs, optimizing storage, and leveraging AI‑driven traffic control, all while sharing practical lessons on monitoring, load testing, and fault‑tolerance.
Lazada transformed its e‑commerce platform for Southeast Asia by fully cloud‑nativeizing core systems, deploying serverless and container technologies, leveraging AI and progressive hybrid apps for low‑end devices, and building robust CDN, promotion, and logistics algorithms that delivered a 99.99% success rate during Double 11.
The article analyzes Baidu's strategic shift toward acquiring YY Live and strengthening its mobile content ecosystem, highlighting how the lack of a unified account system and declining user dwell time have driven the search giant to pursue live‑streaming and recommendation‑driven products to diversify revenue and improve user retention.
This week’s tech roundup covers JD’s unmanned delivery demo, Windows 10’s upcoming Android app support, rapid digital RMB payments, the Kubernetes 1.20 release, China’s IoT security standard, Google’s MediaPipe Holistic for simultaneous face‑hand‑pose prediction, a crystal‑graph neural network for materials science, and emerging trends in multi‑label learning.
This article presents a comprehensive study on automatically extracting theme‑based recommendation reasons for travel content, detailing a three‑stage retrieval framework, the advantages of interactive matching models over classification, rule‑based and back‑translation data augmentation techniques, and various model optimization strategies including priors, transfer learning, seed selection, optimizer choice, and layer‑wise learning rates.
This article explores how AI is being applied to waste classification—from smart trash cans and autonomous garbage trucks to deep‑learning models—while highlighting data‑labeling hurdles, model selection pitfalls, and how platforms like Huawei Cloud ModelArts can streamline development.
iQiyi and Xiamen University have created an AI‑powered Film Restoration Joint Laboratory that merges the university’s film‑history expertise with iQiyi’s video‑understanding and enhancement technology, delivering restoration up to 500 times faster and debuting a 4K, AI‑enhanced version of the 1960s classic “Little Soldier Zhang Ga” while preserving its original artistic style.
iQIYI’s AI‑driven UI element recognizer uses a YOLOv3 model trained on thousands of mobile and PC screenshots to locate obfuscated controls across diverse devices, integrating predictions into its IDE for reliable automation of React Native, Flutter, H5 and mini‑program interfaces.
The article explains how building account relationship graphs and applying both traditional graph algorithms and modern graph neural networks—through community detection, anomaly detection, semi‑supervised and unsupervised learning, and dynamic graph techniques—can effectively identify and dismantle fraud groups in online services.
This roundup covers recent tech headlines—from Alibaba’s NeurIPS‑featured 3D AI model that turns 2D images into searchable 3D objects, to WeBank’s interest‑free rent‑loan extension, Honor’s staff integration plan, a critical iPhone remote‑control flaw, and links to deep‑dive articles on Linux kernel memory handling, command‑line productivity tools, and database cache consistency.
The online AI Engineering Architecture Salon, organized by 58.com AI Lab, will present four technical sessions on December 3 and 8 covering backend design for intelligent customer service, speech recognition engine, the Lingxi voice analysis platform, and a high‑performance vector retrieval system, each with detailed abstracts and expert speakers.
The DataFunTalk year‑end conference will be held online on December 19‑20, featuring over 90 speakers across multiple forums covering recommendation algorithms, knowledge graphs, AI, big data, security, and product development, with detailed session schedules, speaker bios, and registration information.
The latest developer newsletter highlights JD's new Explore research institute, Microsoft's Cloud PC preview, CNCF's etcd graduation, JD's AI-powered customer service brand, a CS‑1 AI supercomputer beating traditional supercomputers, PHP 8.0 release with JIT, curriculum learning for NLU, and advances in 3‑D object reconstruction.
This article introduces NVIDIA's Merlin recommendation system framework—including NVTabular, HugeCTR, and Triton—covers common recommendation pipelines, industry challenges, component advantages, deep‑learning model examples, and performance comparisons, providing a comprehensive guide for building high‑performance AI‑driven recommender systems.
This case study details how a city‑bank leveraged DevOps and ITIL integration, AI‑driven monitoring, and Spark‑based big‑data analytics to build a unified development‑testing‑operations platform, improve service availability, shorten deployment cycles, and achieve near‑99.99% system uptime across its core banking services.
The article proposes a supply‑filtered, tree‑based approach to discover multi‑dimensional user housing preference schemes, contrasting fixed‑length preference mining methods, and details algorithmic modules such as split‑point search, similarity calculation, split suppression, and user clustering to improve interpretability and offline applicability.
iQIYI’s Machine Learning Platform evolved from the specialized Javis deep‑learning system into a unified, low‑threshold solution for algorithm engineers, analysts, and developers, adding visual pipeline building, multi‑framework scheduling, automatic hyper‑parameter tuning, parameter‑server training, and scalable online prediction, dramatically boosting business efficiency and detection performance.
Li Yongxiang, technical director at Suning Tech, outlines how AI-driven visual unmanned stores and integrated big‑data, cloud, and edge computing solutions enable low‑cost digital transformation across thousands of retail outlets, improving shopper experience, inventory management, and operational efficiency.
Meituan’s autonomous vehicle engine abstracts communication, scheduling, data handling, and tooling into three layers to ensure deterministic behavior across on‑vehicle and simulation environments, tackling consistency, scheduling, and resource‑utilization challenges by using unified computation models, distributed graph deployment, caching, and remote model serving, thereby accelerating autonomous delivery vehicle development.
Rem, a former materials engineering student turned front‑end engineer, relentlessly contributes to open‑source, leads the AI‑focused Pipcook project, and shares his philosophy of breaking limits, bridging front‑end development with machine learning, and empowering developers through community‑driven innovation.
This article provides a comprehensive, illustrated walkthrough of OpenAI's GPT‑2 language model, covering its decoder‑only Transformer architecture, self‑attention mechanisms, token processing, training data, differences from BERT, and applications beyond language modeling, enriched with visual diagrams and code snippets for deeper understanding.
JD Retail’s CEO outlines how the company’s AI‑driven supply‑chain, JD Retail Cloud, and the new Shangling SaaS platform are transforming traditional retail, cutting inventory turnover to 34 days, boosting brand sales over 300%, and driving a ten‑fold rise in C2M product launches.
This article presents Baidu's large‑scale event graph technology, detailing its concepts, construction pipeline, detection, representation, extraction methods, causal relation mining, and real‑world applications such as hotspot discovery, event tracing, resource linking, and causal reasoning, while outlining future research directions.
The article reports that China’s AI core industry reached 770 billion yuan in the first half of the year, while also covering BOSS recruitment abuse actions, a Tesla sunroof incident, Huawei’s investment in Quanzhi Microelectronics, Alibaba Cloud’s optimistic view on autonomous driving, and links to deeper technical pieces on Linux, command‑line tools, and database caching.
This roundup highlights 2020's most common passwords, Chrome 87's performance surge, Gartner's cloud spending forecast, Facebook's AI hate‑speech detection, ZTE's AI‑driven energy saving, AMD's MI100 AI accelerator, curriculum‑learning NLU improvements, and EMNLP's silent‑speech breakthrough.
This article introduces NVIDIA's Merlin recommendation system framework, detailing its three core components—NVTabular for feature engineering, HugeCTR for high‑performance CTR model training, and Triton for inference—while discussing common pipeline challenges, performance advantages, and example implementations for deep‑learning‑based recommender models.
This article presents Meituan's AI-driven short video analysis workflow, covering industry trends, multi‑label video classification, intelligent cover selection, and video generation techniques, while discussing challenges, model building, label expansion, continuous data iteration, and future outlook for video AI in local services.
The paper introduces a symbolic‑logic‑focused evaluation set for knowledge‑base question answering, defining single, deterministic, and non‑deterministic element symbols with linguistic tags, constructing complex celebrity‑domain queries, and showing that current commercial KBQA systems achieve only modest accuracy, highlighting the need for more robust reasoning capabilities.
iQIYI’s AI‑driven coloring engine uses a GAN‑based model with attention and semantic reference modules to automate line‑art coloring, delivering consistent styles across frames, cutting coloring time to 0.7 seconds per image and reducing labor costs by about 30% in animation pipelines.
The iQIYI M2VoC Challenge at ICASSP 2021 invites researchers to tackle low‑resource multi‑speaker, multi‑style voice cloning by providing Mandarin datasets, few‑shot and extremely few‑shot tracks with strict data rules, MOS‑based subjective evaluation, and a $9,600 prize pool for top submissions.
This article presents a comprehensive catalog of more than 40 open‑source projects released by Didi, covering AI runtimes, speech and NLP engines, big‑data loaders, middleware, mobile frameworks, frontend UI libraries and various operational tools, each with a brief description and a GitHub link.
This article walks through a hands‑on experiment using Huawei Cloud RDS and Python to convert short videos into audio, extract voiceprints, and match them against a database, demonstrating how speech‑signal processing enables accurate identification of background songs in TikTok‑style clips.
This talk details the design, data pipeline, feature engineering, model evolution, and operational insights of the 58 Tongzhen home feed recommendation system, covering its architecture, localization strategies, recall and ranking models, online learning, and future directions for AI-driven content delivery in the down‑market.
Fluid 0.4 introduces a DataLoad custom resource for declarative data pre‑warming, enhances support for massive small‑file datasets, adds HDFS‑compatible access for Spark and other big‑data frameworks, and enables mixed‑deployment of multiple datasets on a single node, all backed by significant performance gains.
This week’s developer roundup covers JD’s open AI‑driven supply‑chain decision engine, LinkedIn’s Dagli Java machine‑learning library, PayPal’s crypto trading rollout, SpaceX’s Starlink service approval in Canada, AppleWatch’s NightWare nightmare‑interrupt feature, Intel’s new server GPU, and two cutting‑edge GNN research papers on multivariate time‑series forecasting and molecular generation.
This article provides a comprehensive overview of federated learning engineering, covering deployment via Docker containers, the design of training and inference frameworks, key services such as communication, training, model management, and registration, and practical considerations for scaling and reliability in production environments.
This article presents a comprehensive study on slot (entity) recognition and error correction for voice robots, describing the labeling scheme, data annotation, IDCNN+CRF and BiLSTM+CRF models, a pinyin‑based similarity algorithm, and reporting significant accuracy improvements in real‑world deployments.
Alibaba’s 2020 Double 11, rebranded as “Double Nunchaku,” turned the shopping festival into a high‑pressure technology showcase, driving cloud‑on architecture, AI‑enhanced experiences, and a digital‑native business operating system that reshapes e‑commerce innovation.
This article presents a comprehensive overview of Ctrip's multilingual machine translation platform, detailing demand analysis, system architecture, data pipeline, algorithmic innovations such as task‑space fusion and term‑translation interventions, as well as extensive performance optimizations for low‑resource languages.
A Scottish football match turned comedic when an AI camera misidentified a linesman's shiny bald head as the ball, highlighting challenges in automated sports broadcasting and prompting fan backlash over the technology's shortcomings.
The 2020 Tmall Double‑11 event shattered records with a peak of 583,000 orders per second, showcasing Alibaba’s digital‑native business operating system that combines cloud‑native migration, AI, big‑data streaming, real‑time video, and intelligent logistics to sustain the world’s largest traffic surge.
On November 7, 2020, Hulu’s Shulei presented how the company’s video advertising platform employs AI-driven algorithms—including binge‑watch prediction, inventory forecasting with Prophet, and a PID‑controller based real‑time allocation system—to enhance targeting, optimize revenue, and improve user experience.
This week’s tech roundup covers Python’s rise in the TIOBE rankings, a new 5G‑Industrial Internet security whitepaper, the affordable Raspberry Pi 400 keyboard computer, Apple’s upcoming Apple Silicon MacBooks, Snapdragon 875 specs, SpaceX’s GPS III launch, the Kornia differentiable vision library, and a novel sound‑event localization method using multi‑beamforming and multitask learning.
This article introduces the fundamentals of object detection, explains the latest YOLOv4 and YOLOv5 architectures, and details the essential components—including data preparation, regularization, backbone, neck, and prediction innovations—along with label smoothing and advanced loss functions for improved detection performance.
This article details the design, data pipeline, feature engineering, model development, and iterative optimization of the 58 Tongzhen local feed recommendation system, covering business background, user profiling, recall strategies, ranking models such as XGBoost, XDeepFM, and online learning, and future directions.
This article surveys knowledge‑graph reasoning, explaining deductive and inductive reasoning fundamentals, description‑logic and logic‑programming approaches, and modern embedding techniques such as TransE, TransH, TransR and TransD, while highlighting their theoretical bases, practical implementations and recent research progress.
This article presents Meituan's AI‑driven short‑video analysis pipeline for local‑life scenarios, covering industry trends, multi‑label classification, intelligent cover selection, and video generation, and discusses model construction, label‑system expansion, continuous data iteration, and practical applications in restaurant and hotel domains.
