1,504 technical terms and definitions
action recognition video, action recognition, recognition video
Action Recognition Video Introduction & Motivation Action Recognition Video addresses a central problem in vision process control and action-classification-temporal-modeling…
action recognition video, action recognition, recognition video
Action Recognition Video Introduction & Motivation Action Recognition Video addresses a central problem in robotics process control and video-action-temporal-classification…
Video Action Recognition Methods Introduction & Motivation Video Action Recognition Methods addresses a central problem in vision processing and action-classification-accuracy…
ai agents
Action Space is the complete set of allowed operations an agent can execute to affect its environment - It is a core method in modern semiconductor AI-agent planning and control…
checkpointing, gradient, recomputation, memory
Gradient checkpointing and gradient accumulation are the two techniques that let you train a model that does not fit in memory. They attack different halves of the training memory…
implant
Activation anneal is the thermal process that converts an ion-implanted semiconductor from a damaged crystal containing mostly misplaced dopant atoms into a repaired lattice…
process integration
Activation Anneal is thermal treatment that electrically activates implanted dopants and repairs implantation damage - It determines final junction resistivity and transistor…
llm architecture
Activation Beacon is the LLM optimization technique that compresses intermediate activations to reduce memory consumption and latency — Activation Beacon is an inference…
gradient checkpointing, memory efficient training, rematerialization, recompute activation
Gradient Checkpointing (Activation Checkpointing) is the memory optimization technique that trades compute for memory during neural network training by selectively storing only…
recompute
Activation Checkpointing The Memory Problem During training, activations from forward pass must be stored for backward pass: Each layer stores intermediate values For large…
rematerialization
Gradient checkpointing and gradient accumulation are the two techniques that let you train a model that does not fit in memory. They attack different halves of the training memory…
gradient checkpointing, memory optimization
Gradient checkpointing and gradient accumulation are the two techniques that let you train a model that does not fit in memory. They attack different halves of the training memory…
reliability
Activation energy extraction is the parameter extraction process that quantifies temperature sensitivity of a reliability mechanism using stress data - the extracted activation…
gelu, silu swish, activation nonlinearity, neural network activations
Activation functions are the reason depth means anything. Stack a hundred linear layers with no nonlinearity between them and the whole thing collapses algebraically into a single…
relu variants activation, swish gelu activation, activation function properties, learnable activation functions
Activation Function Design is the selection and engineering of nonlinear transformations applied element-wise to neuron outputs — introducing the nonlinearity essential for neural…
activation function design and selection, activation function design, function design selection, activation function, function design, design selection
Activation Function Design and Selection Introduction & Motivation Activation Function Design and Selection addresses a central problem in ml-architecture process control…
fused activation function implementation and latency, fused activation function, activation function implementation, function implementation latency, fused activation, activation function
Fused Activation Function Implementation and Latency Introduction & Motivation Fused Activation Function Implementation and Latency addresses a central problem in inference…
activation function saturation prediction, activation function saturation, function saturation prediction, activation function, function saturation, saturation prediction
Activation Function Saturation Prediction with Machine Learning Introduction & Motivation Activation Function Saturation Prediction with Machine Learning addresses a central…
neural architecture
Activation Function Zoo refers to the large and growing collection of activation functions available for neural networks — from the classic sigmoid and tanh to modern learnable…
nonlinear transformations, relu variants, gelu swish activations, neural network nonlinearities
Activation functions are the reason depth means anything. Stack a hundred linear layers with no nonlinearity between them and the whole thing collapses algebraically into a single…
activation functions non linearity, activation functions, functions non, cifar gelu glu glue nlp, activation functions & non-linearity, functions non-linearity
Activation Functions & Non-linearity Introduction & Motivation Activation Functions: introduce non-linearity. ReLU, Sigmoid, Tanh, GELU. Applications: deep networks, improved…
activation functions, relu sigmoid tanh, bert elu gelu gpt nlp passed, activation functions relu sigmoid tanh & gelu, activation functions relu, functions relu sigmoid, sigmoid tanh gelu
Activation Functions: ReLU, Sigmoid, Tanh & GELU Introduction & Motivation Activation functions: non-linearities enabling deep learning. ReLU: simple, efficient; dead neuron risk.…
ReLU GELU SiLU, function characteristics, gradient properties, modern architecture choices
Activation Functions Survey (ReLU, GELU, SiLU) compares fundamental non-linearities used in deep learning that introduce non-linearity enabling neural networks to learn complex…
activation gradient recomputation vs storage memory trade-offs, activation gradient recomputation, gradient recomputation vs, recomputation vs storage, vs storage memory, vs storage
Activation Gradient Recomputation vs Storage Memory Trade-Offs Introduction & Motivation Activation Gradient Recomputation vs Storage Memory Trade-Offs addresses a central problem…
explainable ai
Activation Maximization is the optimization-based approach to generating inputs that maximally activate a target neuron or output class in a neural network — using gradient ascent…
interpretability
Activation Maximization is optimization of input patterns to maximize a chosen neuron, channel, or class activation - It exposes preferred stimulus patterns encoded by model…
explainable ai
Activation maximization for text is the optimization approach that searches for text inputs which maximize a chosen internal activation in a language model - it is used…
causal, intervention
Activation Patching (Causal Tracing) is the mechanistic interpretability technique that identifies which specific components of a neural network causally store particular…
ai safety
Activation patching edits internal activations to understand the causal role of specific neurons, layers, or circuits. Technique: Run model on two inputs (clean and corrupted)…
explainable ai
Activation patching is the causal intervention method that replaces selected activations in one run with activations from another run to test influence on outputs - it is one…
sparse activation, relu sparsity, conditional computation, sparse inference
Activation Sparsity in Neural Networks is the phenomenon and optimization technique where a large fraction of neuron activations are zero or near-zero during inference — enabling…
interpretability
Activation Steering is a control technique that modifies hidden activations to bias model outputs - It enables behavior shaping at inference time without full retraining. What Is…
interposer, chiplet, integration, routing
Active Interposer Design Integration is a silicon substrate containing embedded logic, routing resources, and power management circuits that actively orchestrates communication…
Active Anomaly Discovery with Human Feedback Integration Introduction & Motivation Active Anomaly Discovery with Human Feedback Integration addresses a central problem in anomaly…
Active Fairness Auditing for Disparate Subgroups Introduction & Motivation Active Fairness Auditing for Disparate Subgroups addresses a central problem in {subject} learning…
Active Class Selection for Controlled Class Proportions Introduction & Motivation Active Class Selection for Controlled Class Proportions addresses a central problem in {subject}…
Active Clustering and Labeling Introduction & Motivation Active Clustering and Labeling addresses a central problem in {subject} learning and efficient {target} annotation: how…
Active Continual Learning with Information-Theoretic Queries Introduction & Motivation Active Continual Learning with Information-Theoretic Queries addresses a central problem…
active cooling control stability prediction, active cooling control, cooling control stability, control stability prediction, active cooling, cooling control, control stability
Active Cooling Control Stability Prediction with Machine Learning Introduction & Motivation Active Cooling Control Stability Prediction with Machine Learning addresses a central…
active cooling fan design, active cooling fan, cooling fan design, active cooling, cooling fan, fan design
Active Cooling Fan Design Introduction & Motivation Active Cooling Fan Design addresses a central problem in cooling process control and cooling-capacity-watts prediction: how…
Active Feature Acquisition for Costly Missing Attributes Introduction & Motivation Active Feature Acquisition for Costly Missing Attributes addresses a central problem…
Active Feature Exploration via Gaussian Process Uncertainty Introduction & Motivation Active Feature Exploration via Gaussian Process Uncertainty addresses a central problem…
Active Few-Shot Learning Introduction & Motivation Active Few-Shot Learning addresses a central problem in fewshot and sample-selection-few-shot architecture: how to Implement…
uncertainty sampling, query by committee, diversity sampling, human annotation
Active learning iteratively chooses which unlabeled examples should receive costly labels to maximize information gained per annotation. It reduces labeling burden when expert…
query strategy active learning, uncertainty sampling, pool based active learning, annotation efficient learning
Active Learning is the iterative machine learning framework where the model itself selects the most informative unlabeled examples to be annotated by a human oracle, minimizing…
active learning annotation, active learning, learning annotation
Active Learning Annotation Introduction & Motivation Active Learning Annotation addresses a central problem in nlp process control and active-learning-sample-selection prediction…
Active Learning for Computer Vision Introduction & Motivation Active Learning for Computer Vision addresses a central problem in {subject} learning and efficient {target}…
Curriculum Active Learning Introduction & Motivation Curriculum Active Learning addresses a central problem in {subject} learning and efficient {target} annotation: how to Achieve…
Ensemble Diversity in Active Learning Introduction & Motivation Ensemble Diversity in Active Learning addresses a central problem in {subject} learning and efficient {target}…
data
Active learning for annotation is a machine learning strategy that intelligently selects which unlabeled examples should be annotated next, focusing human effort on the examples…