The article dissects libstdc++'s std::sort implementation, explaining how introsort combines quicksort, heap sort, and insertion sort, why the recursion threshold of 16 and the depth limit of 2·log₂(n) are chosen, and how a sentinel trick eliminates boundary checks for extra performance.
A recent 154‑page paper co‑authored by Google and top universities reveals how large language models can accelerate frontier scientific research through iterative prompting, cross‑disciplinary knowledge transfer, automated proof verification, and novel algorithmic insights, reshaping the workflow of mathematicians and scientists alike.
This article explains AI compute power as the computer's ability to process data, describes why strong compute accelerates model training, outlines the three main types—general, intelligent, and super—and breaks down its hardware, software, algorithm, and infrastructure components for beginners.
This article analyzes high‑throughput sensitive‑word filtering by comparing brute‑force, KMP, Trie, double‑array Trie and Aho‑Corasick algorithms, presents their time and space complexities, shows Java implementations for Trie and AC automata, evaluates Netty deployment options, and offers practical optimizations such as asynchronous detection, hot‑reloading, tiered responses, logging and fuzzy matching.
This comprehensive guide examines large‑scale deep reinforcement learning, detailing policy‑gradient fundamentals, the mathematics of PPO and GAE, practical implementation tricks, reward and observation normalization, network initialization, and the newer Phasic Policy Gradient method, all supported by code snippets and key research references.
Cloud Music revamped its push‑notification system by separating business and channel layers, integrating a unified delivery platform, tailoring messages to Android manufacturers, adding new push channels, refining frequency and copy controls, and using AI‑generated creatives, which together doubled click‑through rates and nearly doubled total click users within two months.
The article details how Autohome tackled the cold‑start problem for newly generated content by redesigning the recommendation pipeline, introducing multi‑path recall, refining ranking and re‑ranking formulas, and applying strategic controls, resulting in a rise of cold‑start success rate from 27% to over 99% and a CTR increase from 5% to 14%.
This article details Tencent Cloud's V265/TXAV1 live streaming solution, covering its high‑efficiency 8K and low‑latency capabilities, performance gains over X265, MV‑HEVC 3D compression, extensive engineering and algorithmic optimizations, and the resulting speed‑up and quality improvements for ultra‑high‑definition live broadcasts.
Bilibili’s self‑developed VVC encoder, BILIVVC, earned third place in both the 1080p 1 fps and 5 fps tracks of the 2022 MSU Encoder Competition by leveraging extensive VVC‑tool optimizations, fast‑algorithm cooperation, adaptive pre‑analysis and efficient implementation that deliver high quality YUV‑SSIM performance despite its small‑team, one‑year development.
Over the past year, Taobao Live’s video‑quality team, led by former algorithm researcher Xiaocen, built an automated attribution system, real‑time low‑quality detection, and close streamer support, turning pure research into product‑focused engineering that boosted stream quality, trust, and business impact through cross‑domain collaboration.
This article analyzes the execution speed differences between Python's while and for loops, demonstrates benchmark results using timeit, explains the underlying reasons for the performance gap, and shows how built‑in functions and mathematical formulas can achieve dramatically faster computations.
While distributed clusters are popular for big‑data processing, they are not a universal solution; tasks that are hard to partition or involve heavy cross‑node communication often perform better on a well‑optimized single machine, making a careful analysis of workload characteristics essential before scaling out.
This article reviews Alipay's mobile visual algorithm ecosystem, detailing its diverse application scenarios, technical challenges, architectural framework, lightweight design strategies, scalable modeling techniques, and future research directions for edge AI on billions of devices.
This article explains the design and Go‑language implementation of the pattern‑defeating quicksort (pdqsort), a hybrid unstable sorting algorithm that leverages generics and multiple pivot‑selection strategies to achieve 2×‑60× speedups over the standard library in most cases.
The article describes how Vivo's content moderation platform, DiTing, uses algorithm selection—including Aho‑Corasick automaton, combination word matching, and pinyin‑based matching—to efficiently detect sensitive terms in large‑scale text streams, while addressing challenges such as homophones, multi‑character patterns, and performance constraints.
Hello Travel tackles information‑flow advertising challenges by using LightGBM‑based models to predict order conversion, creative performance, and pre‑bid user quality, augmenting sparse data with feature engineering and uplift techniques, while planning future fully automated delivery, richer pre‑screening, and cross‑channel reinforcement‑learning enhancements.
This talk explains how Hello Travel tackles challenges in information‑flow advertising by describing the market landscape, their business background, and detailed algorithmic optimization across plan, creative, and pre‑bid dimensions, including data‑driven modeling, feature engineering, LightGBM and uplift models, and outlines future directions.
This article compares Python's while and for loops, shows benchmark results revealing that for loops run faster due to fewer Python‑level operations, and demonstrates that using built‑in functions like sum or applying a mathematical formula can make looping dramatically faster, often eliminating the loop entirely.
The article thoroughly explains Vue.js’s virtual‑DOM diff algorithm, detailing its depth‑first traversal, same‑level node comparison, the sameVnode key/selector check, index map creation, O(n) head‑tail and index‑based matching loops, Vue 3 static‑type optimizations, and a practical implementation for array change detection.
This article introduces the concept of calibration in trustworthy machine learning, explains why accurate probability estimates are crucial for online advertising, reviews related research and evaluation metrics, and details the evolution of calibration algorithms such as Smoothed Isotonic Regression, Bayes‑SIR, real‑time optimizations, and post‑click conversion models, concluding with engineering deployment and future directions.
The article presents Jump Point Search and four high‑performance variants—JPS‑Bit, JPS‑BitPrune, JPS‑BitPre, and JPS‑BitPrunePre—that combine bitwise acceleration, pruning, preprocessing, and compact multithreaded memory structures to achieve up to 273× faster pathfinding than classic A* on a 200‑grid benchmark.
This article presents a series of algorithmic innovations—including efficient NFA construction, reduced epsilon‑transitions, prefix/suffix optimizations, fast NFA‑to‑DFA conversion, space‑compressed automata, hybrid finite automata, and large‑scale regex matching techniques—designed to improve regular‑expression matching speed and memory usage in deep packet inspection systems.
This article presents a comprehensive study on improving individual diversity in recommendation systems by detailing architectural optimizations across recall, rule, and re‑ranking layers, explaining the principles and practical deployment of MMR and DPP algorithms, and demonstrating their impact on key business metrics through extensive experiments.
The article outlines Qunar’s challenges in aggregating international hotel data, analyzes issues such as localized address formats and limited text similarity parsing, and presents a pattern‑matching and weighted scoring approach that improves aggregation accuracy across multiple countries.
This article walks through three typical interview coding rounds—implementing a robust string‑to‑int parser, designing a high‑concurrency trace‑storage API, and building a matrix multiplication routine with performance and sparse‑matrix optimizations—providing code examples, corner‑case handling, and improvement ideas.
This article explains the quick sort algorithm, covering its basic divide‑and‑conquer principle, a naïve C implementation, improvements such as two‑way partitioning, random and median‑of‑three pivot selection, and a non‑recursive version using an explicit stack, with full source code examples.
This article explores the evolution of recommendation algorithms at Xueqiu, highlighting the limitations of traditional point-wise click-through rate prediction models and detailing the ongoing transition to a pair-wise TF-Ranking framework designed to mitigate user and content biases while significantly enhancing overall recommendation accuracy and user experience.
The department’s extensive 6.18 preparation involved systematic load‑testing, deep learning‑driven personalization of search recommendations, and multiple algorithmic enhancements to improve relevance and conversion, supported by detailed planning, cross‑team coordination, and dedicated night‑shift logistics.
This comprehensive guide explains QuickSort’s divide‑and‑conquer principle, presents two PHP‑style implementations, analyzes its O(N log N) average and O(N²) worst‑case performance, and details practical optimizations such as random pivot selection, switching to insertion sort for small sub‑arrays, three‑way and dual‑pivot variants.
This article describes the evolution of 58.com's user‑set service architecture, the transition from MongoDB to RoaringBitmap storage, and the machine‑learning‑driven algorithm optimizations that enable real‑time, multi‑dimensional, and localized push notifications for millions of users.
The article recounts a interview scenario where a candidate struggled with a basic bubble sort implementation, then explains the standard bubble sort algorithm, its inefficiencies, and demonstrates an optimization that adds a boolean flag to detect a sorted pass and terminate early, improving performance.
Memoization, a dynamic programming technique introduced in 1968, stores results of recursive calls to eliminate redundant calculations, dramatically improving performance—as demonstrated by transforming a naïve O(2ⁿ) Fibonacci function into an O(n) version with simple code modifications and practical examples.
This article explains how Alibaba Xianyu uses GeoHash encoding and optimized spatial indexing to efficiently match billions of user‑posted GPS points with tens of thousands of market‑area polygons, reducing computation from quadrillions to billions of operations through precise point‑polygon algorithms and fast neighbor lookups.
This article explains how Xianyu uses GeoHash‑based spatial indexing, precise and approximate matching, and a rapid neighbor‑search algorithm to efficiently associate billions of GPS‑tagged items with tens of thousands of city‑level business districts, reducing computation from quadrillions to billions of operations.
This article details how Alipay's scanning technology team improved QR code recognition by refining aspect‑ratio tolerance, introducing new pattern detection modes, applying diagonal filtering, leveraging logistic‑regression classification, adjusting jump‑line intervals, and moving binarization to GPU, resulting in a 6.95‑point increase in recognition rate and significantly reduced processing time.
This article introduces internal and external sorting concepts, then walks through Python implementations of bubble sort and selection sort, detailing basic algorithms, optimization techniques such as early‑exit flags and binary selection, and analyzes their time complexities and trade‑offs.
This article demonstrates how to replace the classic linear search loops for arrays and strings with a sentinel‑based approach, reducing boundary checks and improving performance, especially on large datasets, and includes a concrete performance test.
