按方法、领域与证据探索科学。

t-SNE

t-SNE (t-Distributed Stochastic Neighbor Embedding) is a nonlinear dimensionality-reduction method introduced by Laurens van der Maaten and Geoffrey Hinton in 2008 that maps high-dimensional data into a 2D or 3D space for visualization. It preserves probabilistic local similarities, so points that are neighbours in the

deep learning

T5 (Text-to-Text Transfer Transformer)

T5 is a unified sequence-to-sequence deep learning framework introduced by Raffel et al. at Google Brain in 2020, published in the Journal of Machine Learning Research (Vol. 21, No. 140). It reframes every NLP task — classification, translation, summarisation, question answering, and more — as a text-to-text problem: b

cryptography

Taint Analysis

Taint analysis is a data flow analysis technique that tracks how untrusted (tainted) input flows through a program to identify vulnerabilities where tainted data reaches dangerous operations (sinks). Formalized by Newsome and Song in 2005, taint analysis marks input data as tainted and propagates taint labels through t

manufacturing

Taylor Tool Life

Taylor's tool life equation is an empirical relationship predicting how long a cutting tool remains usable before dulling or breaking, expressed as a function of cutting speed, feed rate, and depth of cut. Formulated by Frederick Winslow Taylor in 1907 from systematic experiments on metal cutting, this method provides

aerospace

TCAS

TCAS (Traffic Collision Avoidance System) is an airborne safety system that detects nearby aircraft using radar and mode C altitude reports, then provides traffic advisories (TAs) and recommended collision avoidance maneuvers (RAs) to flight crews. Mandated globally on commercial aircraft since 2000, TCAS is considered

software engineering

Technical Debt Measurement

Technical debt represents accumulated shortcuts, deferred maintenance, and design compromises that incur future costs through slower development, higher defect rates, and deployment difficulty. Introduced by Ward Cunningham (1992), technical debt measurement quantifies these burdens using metrics like code complexity,

numerical methods

Technical Debt Quantification

Technical Debt Quantification is the measurement and monetization of technical shortcuts taken during development (incomplete refactoring, outdated dependencies, deferred testing). Coined by Cunningham in 1992, the metaphor frames accumulated shortcuts as financial debt: taking shortcuts saves immediate time but incurs

music information retrieval

Tempo Estimation

Tempo estimation is the task of automatically determining the beats per minute (BPM) or tempo of a musical recording. Introduced by Scheirer (1998), it is fundamental to rhythm analysis, music classification, and synchronization applications. Tempo is one of the most perceptually salient features of music; accurate est

network analysis

Temporal Betweenness Centrality

Temporal Betweenness Centrality (TBC) extends classical betweenness centrality to time-stamped networks by counting how often a node lies on time-respecting shortest paths — paths that traverse edges in chronological order. It identifies nodes that act as temporal brokers, controlling information or resource flow as it

network analysis

Temporal Closeness Centrality

Temporal closeness centrality extends the classical closeness measure to time-varying networks by replacing static shortest paths with time-respecting (foremost) paths. It quantifies how quickly a node can reach all other nodes when interactions occur at specific moments in time, giving a more realistic picture of info

network analysis

Temporal Community Detection

Temporal community detection identifies cohesive groups (communities) in networks whose structure changes over time. By treating each time snapshot as a network layer and coupling consecutive layers, it reveals how communities form, merge, split, grow, or dissolve — turning a sequence of static snapshots into a continu

network analysis

Temporal Degree Centrality

Temporal degree centrality extends the classic degree centrality to time-varying networks by counting how many distinct contacts a node accumulates over time. Rather than collapsing a dynamic network into a single static graph, it preserves the temporal order of edges, yielding a more faithful measure of a node's activ

network analysis

Temporal Eigenvector Centrality

Temporal eigenvector centrality extends the classical eigenvector centrality to networks that change over time. By accounting for the ordering and timing of connections, it identifies nodes that are influential not merely because of many simultaneous connections, but because they sit at the crossroads of sequentially i

text mining

Temporal Expression Extraction

Temporal expression extraction is a natural-language-processing task that detects dates, times, durations, and frequencies in text and normalises them to the TimeML/TIMEX3 standard. Building on the TempEval shared task introduced by Verhagen et al. (2007), it turns time references scattered through free text into struc

deep learning

Temporal Fusion Transformer

The Temporal Fusion Transformer (TFT), introduced by Lim, Arık, Loeff and Pfister in 2021, is an interpretable deep learning architecture for multi-horizon time series forecasting. It combines variable selection, gating, multi-horizon attention and quantile outputs, processing static, past and known-future inputs toget

network analysis

Temporal Knowledge Graph Analysis

Temporal Knowledge Graph Analysis extends standard knowledge graph methods to data where facts and relationships carry timestamps or validity intervals. It enables reasoning about how entities and relations evolve over time, supporting tasks such as link prediction for future facts, temporal relation classification, an

network analysis

Temporal Modularity Analysis

Temporal modularity analysis extends standard modularity-based community detection to time-varying networks by treating each time slice as a network layer and coupling adjacent layers with inter-temporal links. This allows researchers to identify how communities form, persist, merge, split, and dissolve over time in dy

network analysis

Temporal Multiplex Network Analysis

Temporal multiplex network analysis studies relational systems in which actors are connected by multiple distinct types of relationships that all evolve over time. By simultaneously tracking layer heterogeneity and temporal dynamics, the method reveals how different interaction channels co-evolve, which actors hold per

network analysis

Temporal Network Analysis

Temporal network analysis, formalised by Holme and Saramäki in their landmark 2012 Physics Reports survey, is the study of networks in which edges appear and disappear over time. Rather than collapsing all contacts into a single static graph, the approach preserves the precise timing of interactions — whether as contac

network analysis

Temporal Network Diffusion Analysis

Temporal Network Diffusion Analysis studies how information, disease, influence, or other contagions spread through networks whose structure changes over time. By modeling edges as time-stamped contacts rather than static links, it captures the critical role of timing and ordering in determining which nodes get reached

network analysis

Temporal PageRank

Temporal PageRank extends the classic PageRank algorithm to time-evolving networks by incorporating the recency and ordering of interactions. Edges are weighted by a decay function so that recent contacts contribute more to a node's score than old ones. The result is a dynamic importance ranking that captures who is in

network analysis

Temporal Social Network Analysis

Temporal Social Network Analysis (TSNA) extends classic social network analysis by treating networks as time-varying structures. Rather than aggregating all ties into a single static snapshot, TSNA tracks when ties form, persist, and dissolve, enabling researchers to study how social structures evolve and how dynamic c

network analysis

Temporal Stochastic Block Model

The Temporal Stochastic Block Model (TSBM) extends the classic Stochastic Block Model to sequences of network snapshots, jointly inferring latent community memberships and how those memberships evolve across time. It combines a generative edge-probability model with a Markov process over block assignments, enabling pri

network analysis

Temporal Two-Mode Network Analysis

Temporal two-mode network analysis tracks relationships between two distinct classes of nodes — such as authors and publications, or actors and events — across multiple time points. By combining bipartite structure with longitudinal observation, it reveals how affiliation patterns, collaborations, and community members

aerospace

TERCOM

Terrain Contour Matching (TERCOM) is a terrain-aided navigation method that corrects position estimates by matching altimeter measurements against a stored digital elevation map (DEM). Developed by Boeing in the 1980s for cruise missile guidance, TERCOM enables accurate navigation in GPS-denied environments by exploiti

civil engineering

Terzaghi Consolidation

Terzaghi consolidation theory describes how water-saturated clay soils compress over time as excess pore water pressure dissipates and effective stress increases. Formulated by Karl Terzaghi in 1943, this foundational theory enables prediction of settlement rates for foundations on compressible soils, a critical design

text mining

Text Classification

Text classification, also called text categorization, is a supervised natural-language-processing task that automatically assigns documents to predefined categories. Building on the support-vector-machine approach to text categorization established by Joachims (1998) and consolidated in the text-mining literature by Ag

text mining

Text Coherence Scoring

Text coherence scoring computes a document-level coherence score with machine learning, rooted in the entity-based local coherence model introduced by Barzilay and Lapata (2008). It measures how well the sentences of a text hang together, using either an entity-grid model, a graph-based approach, or a transformer-based

text mining

Text Complexity Analysis

Text complexity analysis measures the linguistic difficulty of a text along dimensions such as syntactic complexity (sentence length, embedded clauses), lexical density, and referential chains. Grounded in readability research consolidated by Vajjala and Meurers (2014) and Crossley and colleagues (2011), it turns prose

text mining

Text Deduplication

Text deduplication is a corpus-quality pipeline that identifies and removes exact and near-duplicate documents from large text collections. Grounded in Andrei Broder's 1997 resemblance theory, it is widely used to improve dataset quality for machine learning model training, search engine indexing, and any downstream NL

text mining

Text Frequency Analysis

Text frequency analysis is a descriptive text-mining method that counts how often words, n-grams, and phrases occur in a corpus to reveal content patterns and dominant themes. It rests on the frequency-distribution insight formalised by George K. Zipf (1949), that a few terms occur very often while most are rare, and i

text mining

Text Infilling

Text infilling is a natural-language-processing task that completes missing words, phrases, or spans in a document by exploiting the surrounding context. Introduced as the cloze procedure by Wilson L. Taylor in 1953 as a readability measure, it was reformulated for neural models by Zhu et al. (2019) and is now used for

text mining

Text Network Analysis

Text network analysis models the words or concepts in a text as nodes and their co-occurrences as edges, then uses network metrics to reveal the structure of meaning. The approach was advanced by Diesner and Carley (2005) for communication networks and by Paranyushkin (2011) for tracing the pathways of meaning circulat

text mining

Text Normalization

Text normalization is an NLP preprocessing pipeline that converts noisy, abbreviated, or misspelled text — such as SMS messages, social-media posts, and OCR output — into a clean, standardised form. It is a prerequisite step for virtually every downstream NLP task, ensuring that inconsistent surface forms do not degrad

text mining

Text Regression

Text-based regression predicts a continuous target variable using features extracted from text — TF-IDF scores, embeddings, or n-grams — as the independent variables. Building on the text-as-data programme consolidated by Gentzkow, Kelly and Taddy (2019), it lets a numeric outcome such as a price, a rating, or a sentim

text mining

Text Segmentation

Text segmentation divides a long document into meaningful sections (segments) along topic or discourse boundaries. Introduced for subtopic passages by Marti A. Hearst's TextTiling (1997), it supports document-structure analysis and the detection of topic transitions in continuous text.

text mining

Text Summarization

Automatic text summarization is a natural-language-processing task that condenses long documents into shorter summaries while preserving their key information. It works through one of two families of approaches — extractive summarization, which selects the most important spans from the source, or abstractive summarizat

deep learning

TextCNN

TextCNN is a convolutional neural network for text classification, introduced by Yoon Kim in 2014, that applies parallel convolution filters of different window sizes over word embeddings to capture local n-gram patterns. It is fast and effective for sentiment analysis and topic classification.

text mining

Textual Entailment

Textual entailment, also known as natural language inference (NLI), is the natural-language-processing task of deciding whether one piece of text (the premise) entails a second piece of text (the hypothesis), contradicts it, or is neutral with respect to it. Formalised by the PASCAL Recognising Textual Entailment Chall

text mining

TF-IDF

TF-IDF, introduced by Salton and Buckley (1988), is a term-weighting scheme that scores each word in a document by how often it appears there and how rare it is across the whole collection. It turns raw text into weighted document vectors, giving high weight to terms that are frequent in one document but uncommon elsew

aerospace

Theodorsen Flutter

Theodorsen flutter analysis is a classical aeroelastic method for predicting the onset of flutter, a self-excited oscillation where aerodynamic forces interact with elastic structural motion to cause rapid growth of oscillations. Developed by Theodore Theodorsen in 1935, the method uses frequency-domain analysis with T

architecture

Thermal Comfort Assessment

Thermal Comfort Assessment is a method for evaluating indoor environmental conditions to predict whether occupants will feel thermally comfortable. Pioneered by Povl Ole Fanger in the 1970s, it combines measurements of air temperature, humidity, air speed, and thermal properties of clothing and activity to determine co

thermodynamics

Thermal Resistance Network

The Thermal Resistance Network method uses electrical circuit analogy to solve heat transfer problems. It treats heat flow as analogous to electric current, thermal resistance analogous to electrical resistance, and temperature difference analogous to voltage potential. This powerful conceptual framework enables engine

meteorology

Thermal Wind

The thermal wind relationship is a fundamental meteorological principle that links vertical wind shear to horizontal temperature gradients. It states that wind speed increases with height in the direction of warming—a direct consequence of hydrostatic and geostrophic balance combined with the ideal gas law.

materials science

Thermogravimetric Analysis

Thermogravimetric Analysis (TGA) is a thermal characterization technique that continuously measures mass loss or gain of a material as a function of temperature (or time at constant temperature). Developed systematically by William Wendlandt and colleagues in the 1960s, TGA identifies thermal transitions (evaporation,

oceanography

Tidal Harmonic Analysis

Tidal harmonic analysis is a mathematical method that decomposes observed sea level or current time series into a sum of sinusoidal components with specific frequencies, amplitudes, and phases corresponding to astronomical tidal constituents. Developed by William Thomson (Lord Kelvin) in 1867, harmonic analysis enables

deep learning

TiDE

TiDE (Time-series Dense Encoder) is an MLP-based encoder-decoder architecture for long-term multivariate time-series forecasting, introduced by Abhimanyu Das and colleagues at Google Research in 2023. The model encodes past time-series observations together with static and dynamic covariates through stacked dense (MLP)

quantum computing

Tight-Binding Model

The Tight-Binding (TB) model is a simplified semi-empirical approach for computing electronic band structures and properties of solids. Formulated by Slater and Koster in 1954, TB treats electron hopping between atomic sites as the dominant interaction, enabling efficient calculations of band dispersion for a wide vari

music information retrieval

Timbre Analysis

Timbre analysis is the computational characterization and modeling of tone color—the perceived quality that distinguishes one instrument from another even at the same pitch and loudness. Pioneered by Grey (1977), timbre analysis extracts acoustic descriptors that characterize spectral shape, temporal dynamics, and harm

quantum computing

Time-Dependent DFT

Time-Dependent Density Functional Theory (TDDFT) extends DFT to excited states and time-dependent phenomena. Formulated by Runge and Gross in 1984, TDDFT enables calculation of excitation energies, optical spectra, and charge-transfer processes with moderate computational cost, making it invaluable for photochemistry a

deep learning

Time-MoE

Time-MoE is a billion-scale autoregressive foundation model for universal time-series forecasting, introduced by Shi et al. in 2024 and accepted at ICLR 2025. It combines a decoder-only transformer architecture with sparse Mixture-of-Experts (MoE) feed-forward layers, enabling the model to scale to billions of paramete

particle physics

Time-of-Flight PID

Time-of-Flight (ToF) particle identification measures the time taken for a particle to travel a known distance, enabling determination of the particle's velocity and mass. This complementary technique to Cherenkov and ionization energy loss provides robust particle separation across wide momentum ranges in modern detec

bioinformatics

Time-series ChIP-seq peak calling

Time-series ChIP-seq peak calling extends standard chromatin immunoprecipitation sequencing analysis to samples collected at multiple time points. By identifying and comparing protein-DNA binding peaks across a temporal dimension, the method reveals how transcription factor occupancy, histone modifications, or chromati

bioinformatics

Time-series copy number variation analysis

Time-series copy number variation (CNV) analysis is a computational genomics pipeline that characterizes chromosomal gains and losses across multiple sequential samples from the same individual or tumor. By comparing copy number profiles at successive time points — such as diagnosis, mid-treatment, relapse — it reconst

bioinformatics

Time-series Epigenome-wide Association Study

A time-series epigenome-wide association study (time-series EWAS) extends the classic cross-sectional EWAS design to longitudinal settings, measuring DNA methylation across the entire epigenome at multiple time points within the same subjects. The goal is to identify CpG sites whose methylation levels change systematic

bioinformatics

Time-series eQTL analysis

Time-series eQTL analysis identifies genetic variants (eQTLs) whose effect on gene expression changes over time or across developmental stages. By combining longitudinal RNA-seq data with individual genotypes, the method captures how the same SNP can activate, silence, or reshape gene regulation at different time point

bioinformatics

Time-series gene set enrichment analysis

Time-series gene set enrichment analysis (TS-GSEA) extends the classical GSEA framework to detect biologically coordinated gene sets — pathways, gene ontology terms, or curated signatures — whose collective expression changes meaningfully over time. Rather than comparing two snapshots, it models the full temporal traje

bioinformatics

Time-series metabolomics analysis

Time-series metabolomics analysis profiles small-molecule metabolites from biological samples collected at multiple, ordered time points, enabling researchers to capture the dynamic flux of metabolic pathways in response to stimuli, disease progression, drug treatment, or developmental change. By integrating longitudin

bioinformatics

Time-series microbiome diversity analysis

Time-series microbiome diversity analysis tracks how the richness, evenness, and community composition of microbial communities change across multiple time points within the same subjects. By combining standard diversity metrics with longitudinal statistical models, it separates true temporal dynamics from inter-indivi

bioinformatics

Time-series pathway enrichment analysis

Time-series pathway enrichment analysis identifies biological pathways whose coordinated gene activity changes significantly across ordered time points. Rather than treating each time point independently, the method models the temporal trajectory of gene expression within each pathway and tests whether entire biologica