The article introduces time‑series data, its ubiquitous examples, the challenges of traditional small models, and proposes a universal time‑series large model that simplifies data preparation and model building, ultimately enabling more efficient and stable industrial AI solutions, now offered as a cloud service.
This tutorial explains the nature of sequential data, why traditional feed‑forward networks struggle with it, and how recurrent architectures such as RNN, LSTM, and GRU capture temporal dependencies, complete with mathematical foundations, training algorithms, and full PyTorch implementations for sentiment analysis, text generation, and encoder‑decoder models.
This article provides a comprehensive overview of natural language processing, tracing its evolution from early symbolic and statistical stages through deep learning breakthroughs, detailing sequence models, key NLP tasks, text representation methods, and the development of modern architectures like RNN, LSTM, GRU, Transformer, and GPT series.
This report details a university big‑data project that built a full pipeline—from raw petrochemical market data and text mining to variable selection, XGBoost/LightGBM/Lasoo regression and RNN/LSTM/GRU models—to predict product prices across multiple horizons, evaluate errors, and deliver an interactive demo.
This article explains how recurrent neural networks (RNNs) add memory to neural models, details the gate mechanisms of LSTM and GRU, compares their structures and parameter counts, and illustrates their use in speech recognition, translation, stock prediction, and video generation, while highlighting practical insights and energy considerations.
This article explains why Transformer models surpass traditional RNN‑based seq2seq architectures by introducing self‑attention, multi‑head attention, and positional encoding, detailing the inner workings of encoders, decoders, and attention mechanisms, and comparing their advantages and limitations across NLP and vision tasks.
This article systematically explores the principles and mathematics behind ARIMA, XGBoost, LSTM, and Transformer models, compares their strengths and weaknesses, and demonstrates a complete end‑to‑end workflow for detecting CPU resource anomalies in a cloud service environment.
This tutorial demonstrates how to apply an LSTM deep‑learning model in Python to forecast stock closing prices, covering data acquisition, preprocessing, model construction, training, evaluation, and visualization of results for time‑series prediction.
This article provides a thorough overview of large language models, explaining the relationship between NLP and LLMs, the evolution from RNN to LSTM, the fundamentals of attention mechanisms, and the architecture and operation of Transformer models, all illustrated with clear examples and diagrams.
This article explains the principles and structures of LSTM networks, introduces the ELMO contextual embedding model with its two‑stage pre‑training and downstream usage, and provides an overview of the Transformer architecture, highlighting their roles in modern NLP tasks.
This article presents technical approaches for generating personalized article titles and automatically synthesizing cover images, covering keyword‑based, click‑sequence‑based, and author‑style‑based title models, as well as image restoration, key‑information extraction, object detection, and layout generation techniques to improve user engagement in recommendation and search systems.
The article describes Ctrip's AIOps approach to improving alert quality by combining statistical methods and machine‑learning models such as LSTM, trend analysis, adaptive threshold calculation, and dynamic‑time‑warping based periodic anomaly detection, achieving significant gains in precision and fault‑recall rates.
This article explains the challenges of processing sequential data, introduces LSTM as a solution to long‑term dependency problems in RNNs, details its cell state and gate mechanisms, showcases its architecture, and provides Python code examples for time‑series forecasting using Keras.
This article presents Ctrip's end‑to‑end solution for precise ad‑inventory forecasting using an LSTM model combined with entity embedding, covering data preprocessing, K‑means clustering, model architecture, offline‑online incremental training, early‑stop mechanisms, evaluation metrics, and Python service deployment.
This article introduces the basic architectures of recurrent neural networks (RNN), LSTM, and GRU, explains PyTorch APIs such as nn.RNN, nn.LSTM, nn.GRU, details their parameters, demonstrates code examples for building and testing these models, and provides practical insights for deep learning practitioners.
This article explains the fundamentals of Recurrent Neural Networks and Long Short‑Term Memory units, their gating mechanisms, and demonstrates a practical Python Keras example that predicts future PM2.5 concentrations using an LSTM model.
The paper proposes clustering thousands of related time series and training separate global LSTM models for each cluster, showing that this reduces heterogeneity, leverages shared information, and improves forecasting accuracy compared to individual models, with extensive experiments on CIF2016 and NN5 datasets.
This article presents a threshold‑free anomaly detection framework for streaming time series that combines an LSTM‑based baseline module with an unsupervised detection module, detailing the architecture, training process, data preprocessing, and experimental results that demonstrate superior accuracy and F1 scores.
This article presents a pandemic‑aware method for predicting national hotel monthly room‑nights over the next six months, detailing data augmentation, feature engineering, LSTM and SARIMA‑LASSO modeling, scenario‑based risk assessment, and evaluation results that demonstrate accurate forecasts despite COVID‑19 disruptions.
This tutorial explains how to build, train, and evaluate a multivariate LSTM model using Keras for hourly air‑pollution forecasting, covering data preparation, model design, prediction, and inverse scaling back to original units.
This article explains the structure and operation of recurrent neural networks (RNNs), their limitations, how long short‑term memory (LSTM) networks overcome these issues with gated mechanisms, and provides a complete Python implementation for time‑series airline passenger forecasting.
This article introduces recurrent neural networks (RNNs) and their ability to handle sequential data, explains the limitations of vanilla RNNs, presents the LSTM architecture with its gates, and provides complete Keras code for data loading, model building, and training both vanilla RNN and LSTM models.
This tutorial walks through loading Google stock data, preprocessing it with scaling, constructing past‑window features, building and tuning an LSTM model using GridSearchCV, evaluating predictions, and finally forecasting future values, all illustrated with complete Python code.
The article describes how a logistics company built a data‑driven platform that uses big‑data storage, DeepFM and LSTM models, and real‑time GPS tracking to create an intelligent vehicle‑cargo matching system, a multi‑label driver tagging framework, and a freight price prediction engine, thereby improving efficiency and reducing costs across the industry.
This article describes how a logistics company built an intelligent, data‑driven platform that uses big‑data pipelines, DeepFM for cargo‑driver matching, LSTM for driver‑tag prediction, and cost‑aware models for freight‑price forecasting, improving efficiency and reducing operational costs.
This article explains a practical, supervised approach to concept‑drift detection by labeling historical data as 0 and new samples as 1, training an LSTM (or LightGBM) classifier, using high‑confidence predictions to flag distribution‑inconsistent samples, and includes code and real‑world examples.
This tutorial explains how to work with time‑series data in Python, covering dataset preparation, resampling techniques, Prophet forecasting, LSTM‑based prediction, and hierarchical clustering, while illustrating each step with visual examples and code snippets.
This article surveys deep‑learning image captioning, detailing the image encoder, sequence decoder, attention mechanisms and multimodal designs, comparing encoder‑decoder, detection‑backbone, transformer and dense captioning architectures, and explaining generation strategies and BLEU evaluation.
This article examines how financial institutions can manage risk and improve operations for existing loan customers by analyzing customer flow during the pandemic, policy impacts, rapid deterioration patterns, and by applying advanced models such as LSTM, survival analysis, and B‑card strategies to enable timely risk detection and targeted cross‑selling.
This article presents a practical study of extracting phone numbers from two‑speaker voice dialogues using Seq2Seq models—including LSTM, GRU with attention and feature fusion, and Transformer—detailing data characteristics, model architectures, training strategies, experimental results, and comparative analysis showing the GRU‑Attention approach achieving the best performance.
This article details the BangBang intelligent customer service system's overall architecture, core capabilities, knowledge‑base construction, and successive model upgrades—from FastText to TextCNN, Bi‑LSTM, and model fusion—showing how each iteration improved accuracy, recall, and F1 scores toward a stable 95% performance level.
This article introduces the concept of user preference mining, discusses challenges such as accurate expression, interpretability, and high-dimensional preferences, reviews statistical and model-based approaches including weighting, decay, XGBoost, DNN, LSTM, Seq4Rec, and Deep Interest Network, and describes their practical implementation at Beike.
This article details the 2020 Tencent Advertising Algorithm Competition, describing the user profiling task, data fields, feature engineering, Python code for ID mapping and Word2Vec training, multiple model architectures (LSTM, CNN-Inception, transformer), and the final performance results achieved by the team.
This article explains how WeBank’s intelligent operations team combines LSTM‑based forecasting with Gaussian analysis, k‑means feature clustering, and probability‑density modeling to automatically detect and warn about anomalies in key business metrics, moving beyond traditional threshold‑based monitoring.
The article surveys fine‑grained character sentiment analysis for script evaluation, detailing traditional, target‑dependent and aspect‑level methods, describing iQIYI’s BERT‑TD‑LSTM and CNN architectures, addressing challenges such as character name recognition and long‑range context, and outlining future improvements after a Parasite case study.
This tutorial walks you through using Pipcook’s text‑creation pipeline to train a TensorFlow.js LSTM model that generates Chinese poems from a given seed, covering plugin configuration, GPU training, model evaluation, and Node.js inference.
This article presents an intelligent inventory management solution that predicts product consumption using two time‑series algorithms—Facebook's Prophet and LSTM deep learning—detailing data sources, preprocessing, model configuration, evaluation metrics, and a comparative analysis of their performance and suitability.
The article proposes a deep‑learning framework that uses an LSTM to predict segment‑level travel times and fully‑connected layers to aggregate them into a full‑route ETA, demonstrating on Beijing data a 28.2% MSE reduction and superior accuracy over traditional regressors by capturing temporal and network dependencies.
This article describes how the 58 used‑car team built a speaker‑verification pipeline—covering data collection, MFCC feature extraction, LSTM and GMM modeling, threshold tuning, multi‑speaker clustering, and deployment results—to automatically block nuisance telemarketing calls while preserving user privacy.
This article explains knowledge tracing and its deep learning variant DKT, detailing the underlying RNN/LSTM models, data encoding, loss functions, experimental results, practical applications in adaptive learning, as well as advantages, limitations, and future research directions.
iQIYI’s brand‑advertising system combines statistical and ARIMA‑based forecasting, adaptive and deep‑learning models, factorization‑machine regression, large‑scale bipartite‑graph allocation, hierarchical handling of long‑tail dimensions, frequency‑capping constraints, and an AI‑driven video‑tagging pipeline to accurately estimate inventory and dynamically place ads.
This article describes how Gaode Map leverages machine‑learning techniques—such as word2vec embeddings, LSTM networks, fine‑tuning, and confidence‑threshold ensembles—to automatically classify and verify massive user‑feedback intelligence, streamlining the multi‑step workflow from data collection to road‑map updates and dramatically improving efficiency.
This article details JD Logistics' journey to design and implement a massive, AI‑enhanced monitoring platform that handles over three million metrics across hundreds of warehouses, addressing challenges of scale, network complexity, frequent asset changes, and integrating AIOps for proactive fault detection and resolution.
This article details a deep‑learning‑driven, real‑time alert system that combines TensorFlow LSTM time‑series forecasting with Spark Streaming to achieve high‑coverage, low‑latency anomaly detection for large‑scale data‑ops environments, including data preprocessing, metric classification, model training, and deployment pipelines.
This article describes the TPNN deep‑learning platform’s multi‑GPU acceleration, data‑parallel BMUF training, LSTM‑CTC acoustic modeling, and a suite of mobile‑side optimizations—including model pruning, 8‑bit quantization, low‑precision matrix multiplication and mixed‑precision computation—that together achieve over 92% recognition accuracy for children’s English speech on both server and mobile devices.
This article investigates whether a short‑term time‑series algorithm, specifically an LSTM model trained on limited SARS data, can predict and assess COVID‑19 case numbers, describing data collection, model training, experimental validation, error analysis, and practical implications of the findings.
Amap replaced its rule‑based feedback pipeline with a three‑stage, LSTM‑driven system that combines word2vec embeddings and structured fields, achieving over 96% classification accuracy, cutting manual workload by 80%, and slashing per‑task costs while enabling scalable, data‑driven map quality improvements.
The article reviews traditional, neural network, and open‑source time‑series forecasting methods, explains their strengths and limitations, and demonstrates how NetEase applies short‑term and long‑term prediction models such as Holt‑Winters, ARIMA, STL, Prophet, and LSTM to improve game monitoring and proactive alerting.
The iQIYI advertising team introduced an LSTM‑based deep‑learning model that forecasts inventory by normalizing data, clustering dimensions, and embedding fine‑grained holiday features, achieving significantly lower bias than their Adaptive‑ARIMA baseline and improving generalization while reducing training resources.
This article provides a comprehensive overview of deep learning fundamentals, covering neural network basics, forward and backward feedback architectures, key models such as MLP, CNN, RNN, LSTM and GRU, training techniques like gradient descent, learning rate schedules, momentum, weight decay, and batch normalization.
Amap's POI naming automation uses a two-stage cascade model: Stage 1 extracts token and sentence features with POS tags and domain-adapted BERT‑POI; Stage 2 employs a Bi‑LSTM to model line relationships, achieving over 95% semantic accuracy and 3‑6% recall improvements, thereby enhancing automatic signboard‑based POI name generation.
StarNet is a neural network for pedestrian trajectory prediction in large‑scale delivery, using a global dynamic map and a Hub‑Host architecture to model interactions efficiently, reducing complexity from O(N²) to O(N), and achieving higher accuracy with fast inference compared to baseline methods.
Meituan’s unmanned‑delivery team won the CVPR 2019 Trajectory Prediction Challenge by using a multi‑class independent LSTM encoder‑decoder with Gaussian‑noise augmentation, discarding size and noisy orientation data, applying rotation and interpolation augmentations, and achieving a weighted ADE of 1.3425, surpassing StarNet and TrafficPredict, with plans to explore interaction‑based and graph‑neural‑network models.
This article presents a comprehensive study of near‑duplicate video retrieval, covering the definition of near‑duplicate videos, motivations for deduplication, challenges, a two‑stage offline/online processing framework, keyframe and VGG16‑based feature extraction, metric‑learning loss functions, training procedures, dataset preparation, evaluation metrics, and model enhancements using LSTM and attention mechanisms.
The article details a practical workflow for Chinese sentiment classification in Tencent’s Goose Man product, covering data preparation, word‑segmentation challenges, a six‑layer multi‑LSTM architecture with word embeddings, training results achieving roughly 96 % accuracy, and its deployment for automatic detection of misleading and high‑impact user reviews.
This article reviews the fundamentals of automatic speech recognition, compares traditional MFCC‑GMM‑HMM pipelines with modern deep neural network approaches such as DNN‑HMM, LSTM‑CTC, and attention‑based models, and illustrates each evolution step with flowchart diagrams and key references.
This article explains how Long Short‑Term Memory (LSTM) networks overcome the long‑term dependency problem of traditional recurrent neural networks by using gated cells that store and retrieve information over extended periods, and highlights their widespread applications in speech recognition, machine translation, image captioning, and beyond.
This article explains the concepts, training pipeline, deployment formats, and practical applications of language models—particularly LSTM‑based models—for evaluating and generating text quality in a real‑world rental listing platform, highlighting data preparation, model training, and online serving techniques.
This article presents a detailed case study of building a purchase‑user prediction model by integrating Convolutional Neural Networks for feature extraction with Long Short‑Term Memory networks for time‑series forecasting, covering background, model structure, data augmentation, experimental results, and business impact.
This article explains how to use machine‑learning techniques—including data preprocessing, N‑gram, CBOW, and various RNN/LSTM models—to automatically recommend the next function in a test‑case step sequence, improving writing speed and efficiency for developers.
This article describes how a large‑scale operations monitoring system at Ctrip evolved from rule‑based alerts to a deep‑learning‑driven anomaly detection pipeline using LSTM models and Flink streaming, achieving a ten‑fold reduction in alert volume while improving fault detection.
This article introduces Advanced LSTM (A‑LSTM), which linearly combines multiple past hidden states to overcome traditional LSTM's one‑step dependency, and demonstrates its application in utterance‑level speech emotion recognition, achieving a 5.5% accuracy improvement through attention‑based weighted‑pooling RNNs and auxiliary speaker and gender tasks.
The article details Weibo Advertising's Hubble monitoring system, describing its three‑layer architecture, metric taxonomy, AI‑driven trend prediction with LSTM models, dynamic alert thresholds, and performance testing using GoReplay, illustrating how large‑scale data and machine learning enable proactive operations.
This article describes Tubi's experiment with a Long Short‑Term Memory (LSTM) based recommender, covering the model architecture, training setup, performance metrics, and observed recommendation improvements on their video streaming platform.
This article explains how Alipay combats the surge of mobile payment fraud by upgrading its risk‑identification system with deep‑learning techniques, modeling victim and fraudster behavior sequences using LSTM, and integrating the resulting scores into existing models to achieve a measurable increase in detection coverage.
JD AI Research Institute presented a GCN‑LSTM encoder‑decoder system that integrates object semantic and spatial relationships via graph convolutional networks to significantly improve image captioning performance on the COCO benchmark, achieving state‑of‑the‑art results.
This article explains how LSTM networks overcome RNN limitations by using input, forget, and output gates with sigmoid and tanh activations, describes the core update equations, discusses alternative activation functions and hard‑gate variants, and provides references for deeper study.
This article explains what Recurrent Neural Networks are, how they create text representations, why they suffer from gradient vanishing or explosion, and what architectural improvements such as LSTM, GRU, residual connections and gradient clipping can do to overcome these issues.
This article explains how Baidu Waimai's risk‑control team uses recurrent neural networks, especially LSTM, within TensorFlow to detect fraudulent merchants and users, compares static and dynamic RNN implementations, demonstrates a MNIST digit‑recognition example, and discusses optimization algorithms and model trade‑offs for real‑time fraud detection.
An Alibaba engineer created MusicGo, an AI program that scrapes hip‑hop lyrics, trains a seq2seq LSTM model, and generates rap songs themed around Double 11 and Alibaba's smart logistics, illustrating the practical steps, challenges, and creative adjustments needed for AI‑driven music creation.
This article demonstrates how to build and train a bi‑directional LSTM sentiment analysis model with TensorFlow, covering data preprocessing, word embedding, model architecture, code implementation, and evaluation results, and discusses potential improvements.
The paper introduces LSTM‑Jump, a reinforcement‑learning‑trained LSTM variant that can dynamically skip irrelevant tokens, achieving up to six‑fold speed‑ups over standard sequential LSTMs while maintaining or improving accuracy on various NLP tasks such as sentiment analysis, document classification, and question answering.
This article provides a comprehensive guide to Long Short-Term Memory networks, covering their motivation, detailed forward and backward equations, gate mechanisms, training algorithm, gradient checking, and a full C++ implementation, while also introducing the simpler GRU variant.
