Tri-Training
semi-supervised learning
Tri-Training is a highly robust, semi-supervised machine learning algorithm that significantly improves upon standard self-training by utilizing an ensemble of three independent…
27,049 technical terms and definitions
semi-supervised learning
Tri-Training is a highly robust, semi-supervised machine learning algorithm that significantly improves upon standard self-training by utilizing an ensemble of three independent…
advanced training
Tri-training is a semi-supervised approach where three classifiers iteratively label data for each other - Pseudo-label acceptance uses disagreement patterns to reduce individual…
Tri-Training with Oracle Querying on Dual Disagreement Introduction & Motivation Tri-Training with Oracle Querying on Dual Disagreement addresses a central problem in {subject}…
Tri-Training with Two-to-One Disagreement Routing Introduction & Motivation Tri-Training with Two-to-One Disagreement Routing addresses a central problem in semisupervised…
esd
Electrostatic Discharge (ESD) protection constitutes the dedicated on-chip network of high-current shunting devices engineered to safeguard sensitive gate dielectrics, thin tunnel…
trie string performance, trie string, string performance
Trie String Performance Introduction & Motivation Trie String Performance addresses a central problem in data-structures process control and search-time-string-length prediction…
environmental & sustainability
Trigeneration is combined production of electricity, heating, and cooling from one integrated energy system - It extends cogeneration by converting recovered heat into chilled…
design
Trigger voltage is the voltage threshold at which an ESD protection clamp activates and begins conducting current to protect sensitive internal circuits — representing…
audio & speech
Triggered Attention is an ASR decoding strategy where attention is activated by external alignment or trigger signals - It stabilizes streaming recognition by restricting…
Trigonometry, Vectors, Calculus, and Rates of Change 1. A Connected View of Motion and Change Trigonometry describes angles and cycles. Vectors describe magnitude and direction.…
trim and form lead coplanarity prediction with machine learning, trim form lead coplanarity prediction machine learning, trim form, lead coplanarity prediction, avi
Trim-and-Form Lead Coplanarity Prediction with Machine Learning Introduction & Motivation Trim-and-Form Lead Coplanarity Prediction with Machine Learning addresses a central…
trim form lead bend accuracy prediction machine learning, trim form, lead bend accuracy, trim form lead bend accuracy prediction, trim form lead, form lead bend, form lead
Trim Form Lead Bend Accuracy Prediction with Machine Learning Introduction & Motivation Trim Form Lead Bend Accuracy Prediction with Machine Learning addresses a central problem…
federated learning
Trimmed Mean is a Byzantine-robust aggregation rule for federated learning that removes the highest and lowest values for each gradient coordinate, then averages the remaining…
Triple Classification in Knowledge Graphs Introduction & Motivation Triple Classification in Knowledge Graphs addresses a central problem in graph network representation…
nlp
Triple Extraction is the NLP technique for extracting subject-predicate-object triples from text to structure information — Triple Extraction transforms unstructured text into…
process integration
Triple Well is an isolation scheme using deep n-well structures to embed independently biased p-well regions - It improves substrate-noise isolation and body-bias flexibility…
process, deep n-well, isolated p-well, body biasing, noise isolation, twin-well, well architecture, analog isolation, substrate noise
Twin-well and triple-well are two generations of the same basic idea — giving NMOS and PMOS transistors their own separately-doped regions of silicon to live in — with triple-well…
process
Triple-Well Technology is a CMOS process option that adds a Deep N-Well (DNW) beneath the standard P-well — creating an electrically isolated P-well "tub" that can be…
computer vision
Triplet Attention is a lightweight attention mechanism that computes cross-dimension interactions between channel and spatial dimensions — using three parallel branches to capture…
margin, distance
Cosine Similarity Overview Cosine Similarity is the most common metric used to measure how similar two documents (vectors) are, irrespective of their size. It measures the cosine…
Triplet Loss and Metric Learning in Graphs Introduction & Motivation Triplet Loss and Metric Learning in Graphs addresses a central problem in graph network representation…
triplet loss metric learning and distance embedding, triplet loss metric, loss metric learning, metric learning distance, learning distance embedding, triplet loss, loss metric
Triplet Loss Metric Learning and Distance Embedding Introduction & Motivation Triplet Loss Metric Learning and Distance Embedding addresses a central problem in triplet loss…
Triplet Loss in Multimodal Learning Introduction & Motivation Triplet Loss in Multimodal Learning addresses a central problem in multimodal learning and triplet-ranking-accuracy…
inference server, serving
Triton Inference Server What is Triton? NVIDIA Triton Inference Server is a production-grade serving platform that supports multiple frameworks, dynamic batching, and GPU…
openai, kernel, python, jit, autotune, fusion
Triton is OpenAI's open-source language and compiler for writing GPU kernels in Python. It sits in the gap between calling a black-box library like cuBLAS and hand-writing CUDA…
infrastructure
Triton is OpenAI's open-source language and compiler for writing GPU kernels in Python. It sits in the gap between calling a black-box library like cuBLAS and hand-writing CUDA…
nvidia, serving
Triton Inference Server is NVIDIA's production inference platform supporting multiple model formats, dynamic batching, model ensembles, and GPU scheduling for high-throughput…
model serving, inference serving framework, mlops serving, model deployment gpu
Triton Inference Server is the open-source model serving framework developed by NVIDIA that provides a production-grade HTTP/gRPC inference endpoint for deploying multiple ML…
openai triton, triton dsl, gpu kernel dsl, triton compiler
Triton is OpenAI's open-source language and compiler for writing GPU kernels in Python. It sits in the gap between calling a black-box library like cuBLAS and hand-writing CUDA…
single, random
TrivialAugment is the simplest possible automated augmentation strategy that matches or exceeds complex learned policies — applying exactly one randomly selected transformation…
data augmentation
TrivialAugment is an extremely simple augmentation strategy that applies a single randomly selected augmentation with a random magnitude to each image — with zero hyperparameters…
evaluation
TriviaQA is a large-scale reading comprehension dataset containing over 650k question-answer-evidence triples — derived from trivia enthusiasts' websites, it features complex…
evaluation
TriviaQA is a large-scale question answering benchmark derived from trivia questions linked to evidence documents - It is a core method in modern AI evaluation and governance…
rlhf, training
TRL (Transformer Reinforcement Learning) is a Hugging Face library that provides the complete training pipeline for aligning language models with human preferences — implementing…
interpretability
Trojan Attacks on neural networks are attacks that modify the model's weights or architecture to embed a hidden malicious behavior — unlike data poisoning (which modifies training…
ai safety
Trojan Attacks on neural networks are attacks that modify the model's weights or architecture to embed a hidden malicious behavior — unlike data poisoning (which modifies training…
Trojan Neural Network Detection Introduction & Motivation Trojan Neural Network Detection addresses a central problem in robustness of advrobust models under adversarial…
Trojan Neural Network Detection (Extended Study) Introduction & Motivation Trojan Neural Network Detection (Extended Study) addresses a central problem in robustness of advrobust…
Trojan Neural Network Detection (Extended Study) Introduction & Motivation Trojan Neural Network Detection (Extended Study) addresses a central problem in robustness of advrobust…
troubleshooting guide, diagnostic troubleshooting, equipment troubleshooting, fault troubleshooting
Troubleshooting semiconductor manufacturing equipment is the controlled conversion of an observed symptom into a verified cause, a safe repair, and demonstrated restoration…
trpo, reinforcement learning advanced
TRPO is a policy-optimization method in reinforcement learning that constrains updates within a trust region - The algorithm limits policy shift per update, often via…
Trust Region Policy Optimization (TRPO) Introduction & Motivation Trust Region Policy Optimization (TRPO) addresses a central problem in reinforcement policy learning…
true random number generator quality and throughput, true random number, random number generator, number generator quality, generator quality throughput, true random
True Random Number Generator Quality and Throughput Introduction & Motivation True Random Number Generator Quality and Throughput addresses a central problem in security process…
feedback, eval
TruLens is an open-source library for evaluating and tracking LLM applications using the RAG Triad framework — providing feedback functions that score context relevance…
generative models
Truncation Trick is a sampling technique for GANs that improves the visual quality and realism of generated samples by constraining the latent vector to lie closer to the center…
baseten, package
Truss: Model Packaging & Deployment Overview Truss is an open-source framework (by Baseten) for packaging AI/ML models. It solves the "it works on my machine" problem for ML…
recommendation systems
Trust-Based Recommendation is recommendation methods that weight signals using explicit or inferred trust relationships - It prioritizes information from trusted users to improve…
Powell Dogleg Trust-Region Non-Linear Optimization Introduction & Motivation Powell Dogleg Trust-Region Non-Linear Optimization addresses a central problem in optimization…
Trust Region Policy Optimization Introduction & Motivation Trust Region Policy Optimization addresses a central problem in reinforcement policy learning…
trpo, reinforcement learning
TRPO (Trust Region Policy Optimization) is a policy gradient RL algorithm that constrains policy updates to a trust region — ensuring that each update doesn't change the policy…