structured perceptron, structured prediction
Structured perceptron extends the perceptron algorithm to structured prediction by updating on structured outputs rather than class labels.
9,967 technical terms and definitions
Structured perceptron extends the perceptron algorithm to structured prediction by updating on structured outputs rather than class labels.
Structured pruning removes entire channels layers or blocks enabling hardware-efficient acceleration.
Structured pruning removes entire channels or heads. Hardware-friendly. Less flexible than unstructured.
Remove entire channels layers or attention heads.
Represent data with explicit structure.
Structured Support Vector Machines extend SVM framework to structured outputs using max-margin training with structured loss functions.
Stuck-at faults model defects where circuit nodes are permanently fixed at logic zero or one independent of circuit inputs.
Signal permanently stuck at 0 or 1.
Stuck-open faults represent transistors failing to conduct detected through elevated IDDQ or two-pattern testing.
Student is smaller model learning from larger teacher. Teacher provides richer signal than hard labels.
EMA teacher for student training.
AI tutoring explains concepts. Personalized pace, examples.
Self-Test Using MISR and Parallel SRSG provides BIST architecture.
Match style statistics.
Combine styles from different images.
Style mixing combines latent codes at different scales creating hybrid generations.
Match style of reference image.
Style transfer in diffusion models adapts content to reference styles through conditioning.
Apply artistic style from one image to content of another.
Generate content in specific artistic styles.
Style instructions control output: formal/casual, brief/detailed, bullet/prose. Customize communication.
Style-based generator.
Large-scale StyleGAN.
High-quality GAN with style control.
StyleGAN3 improves alias-free generation through rotation and translation equivariance.
Mechanical surface roughness measurement.
Break into multiple queries.
Subgoals are intermediate objectives required to achieve overall goals.
Subgradient methods optimize structured prediction objectives by computing subgradients of non-smooth loss functions.
Determine if one graph contains another.
Subgroup frequency determines sampling rate for detecting process shifts.
Subgroup size affects control chart sensitivity and capability estimation accuracy.
Subject-driven generation creates images featuring specific subjects from reference images.
Subsampling randomly selects training examples reducing privacy cost.
Align domain subspaces.
Atom replacing lattice atom.
Unwanted signals through substrate.
Operate below threshold voltage.
Subtractive etch patterns metal by depositing then etching rather than damascene filling.
Statistical inference from zero failures.
Demonstrate reliability goal.
Successive inspection verifies output before next operation begins.
Successor features decouple environment dynamics from rewards enabling transfer across tasks with shared state space.
Successor representation separates environment dynamics from rewards enabling fast policy evaluation under new rewards.
Suggestion systems encourage employee ideas for improvement.
Summarize old messages to save space.
Summarization compresses long text to key points. Use for fitting long docs into context or creating memory.
Summarize long documents. Key points extraction.
Generate summaries during pre-training.
Stepwise refinement for text generation.