0 views collected around this technical thread.
The article introduces the “System Wind Rose” – a visual model that places a central object at the core and maps multi‑directional influences with varying strengths, probabilities, and dynamics, offering a strategic lens for risk identification, structural balance, and decision‑making across diverse scenarios.
This article explains the core concepts and methods of sensitivity analysis, distinguishes local and global approaches, outlines the simple variation method, and demonstrates how to construct and interpret a tornado chart—complete with a real‑world example of market factors affecting sales—providing clear guidance for robust model evaluation.
The article uses a real‑world raid‑rebuild incident to illustrate why operations teams must understand change background, schedule, and risk, act as project managers, follow a formal change process, and treat production environments with utmost respect.
The 2024 Generative Large Model Security Assessment White Paper, jointly authored by the Chinese Academy of Sciences, the Ministry of Public Security's Third Research Institute, and Ant Group's Ant Security Lab, was unveiled at the inaugural CCF China Data Conference, offering a comprehensive review of model risks, ethical concerns, and evaluation methods to guide research, industry practice, and policy making.
The Beta distribution is a flexible probability model defined on the interval [0,1] with two shape parameters that control its form, offering useful properties such as mean and variance, and is widely applied in A/B testing, risk assessment, and machine‑learning tasks to model proportions and uncertainties.
This article explains how to differentiate between predictive "will it happen" questions and normative "should it happen" decisions, outlines step‑by‑step methodologies for each, and shows how combining prediction and decision‑making can lead to more rational outcomes in science, policy, and business.
The article reviews typical code‑coverage goals such as 70‑80% for system testing, explains how factors like failure cost, resource constraints, testability, and development stage influence target selection, discusses why full coverage is often impractical, and summarizes major industry standards and coverage metrics used in safety‑critical domains.
Statistics permeate daily life, from news to personal decisions, yet common pitfalls like misleading averages, ambiguous percentages, and false causal links often trick us, so understanding these traps helps us interpret data more accurately and avoid faulty judgments.
This article explains how Monte Carlo simulation can be used for sensitivity analysis in portfolio risk assessment, walks through a Python implementation, and demonstrates how varying asset allocations impacts expected return and volatility.
This article explains the concept and calculation of odds ratios, interprets their values, and illustrates their use with examples from epidemiology and marketing, while also discussing their limitations and proper interpretation.
This article details Qunar's DBA team's systematic analysis of shortcomings in their original database inspection and alarm systems, the design and implementation of comprehensive metric enhancements, risk‑level classification, automated reporting, and alarm noise reduction, and reports the significant improvements in stability, efficiency, and fault‑free operation achieved through these optimizations.
The article presents Google DeepMind's AGI evaluation framework, outlining six guiding principles, nine representative definitions, and a hierarchical five‑level classification system to assess AGI performance, autonomy, and societal impact, aiming to provide a common language for model comparison, risk assessment, and progress tracking.
This article explains the Widmark formula, uses Python code and data tables to estimate how long it takes for blood alcohol concentration to fall below the legal driving limit after drinking a 388 ml coffee with alcohol, and explores the impact of weight, gender, and sip size on safety.
An in‑depth exploration of the fuzzy comprehensive evaluation method demonstrates how to convert qualitative safety‑psychology assessments of construction workers into quantitative scores, detailing factor selection, weight determination, membership functions, and a full case study that guides risk‑aware decision‑making.
Narrative design replaces visual mock‑ups with text‑based stories to explore user flows, clarify concepts, assess risks, and gather feedback across platforms, enabling faster iteration and clearer communication between designers, developers, and stakeholders.
This article explains how test engineers can adopt project‑management awareness throughout requirement review, design review, scheduling, test‑case creation, coding, code review, smoke, functional, integration, stability, UAT and regression testing to proactively identify risks, coordinate with stakeholders, and ensure high‑quality, efficient delivery.
This article details the evolution, architecture, and key technologies of Ant Group's anti‑intrusion platform, explaining how it handles trillion‑level data streams for intrusion detection, performs multi‑dimensional risk assessment and attribution, and enables rapid, automated security incident response across massive enterprise environments.
This article presents an end‑to‑end intelligent testing framework that mines development‑stage risk dimensions, conducts admission risk assessment, extracts multi‑dimensional activity data such as coverage metrics, and applies model‑based risk evaluation to guide quality‑assurance decisions and improve release safety.
The article describes an intelligent test‑evaluation framework that gathers performance data, quantifies project, personnel, and code risk dimensions, feeds them into rule‑based and logistic‑regression models to produce risk scores and risk‑driven testing plans, and demonstrates how this approach identified thousands of high‑risk projects, prevented hundreds of bugs, and saved thousands of person‑days.
This article presents fifteen practical guidelines for managing software dependencies, covering pre‑inclusion checks such as design review, code quality, testing, security, licensing, and transitive dependencies, as well as post‑use practices like encapsulation, isolation, update strategies, and continuous monitoring to maintain a secure and reliable supply chain.