Group-Normalization Machine Learning
Group Normalization Variants Introduction & Motivation Group Normalization Variants addresses a central problem in vision processing and small-batch-stability prediction: how…
1,016 technical terms and definitions
Group Normalization Variants Introduction & Motivation Group Normalization Variants addresses a central problem in vision processing and small-batch-stability prediction: how…
recommendation systems
Group Recommendation is recommendation for multi-user groups instead of single-user personalization - It aggregates member preferences to rank items acceptable to the group as…
Group Recommendation Systems Introduction & Motivation Group Recommendation Systems addresses a central problem in recsys systems and {target} optimization: how to Build {subject}…
leak, prevent
GroupKFold is a cross-validation strategy that prevents data leakage by ensuring all samples from the same "group" stay together in either the training set or the test set, never…
computer vision
Grouped Convolution is a convolution where input channels are divided into groups, and each group is convolved independently — reducing parameters and FLOPs by a factor of while…
model optimization
Grouped Convolution is a convolution method that partitions channels into groups processed by separate filter sets - It reduces parameters and compute while preserving…
gqa, kv
Multi-head attention (MHA), multi-query attention (MQA), and grouped-query attention (GQA) are three ways to wire the key and value projections of a Transformer's attention layer.…
gqa, multi query attention, mqa, attention head sharing
Grouped-Query Attention (GQA) is the attention architecture variant that shares Key and Value heads among groups of Query heads — reducing the KV cache memory footprint…
grouped-query attention, gqa, llm architecture
Grouped-Query Attention (GQA) is an attention architecture that provides a tunable middle ground between Multi-Head Attention (MHA) and Multi-Query Attention (MQA) — using G…
multi query attention mqa, kv cache reduction, attention head grouping, llama 2 attention
Grouped Query Attention (GQA) is the attention mechanism that shares key and value projections across groups of query heads, interpolating between multi-head attention (MHA)…
optimization
Grouped-query KV cache is the attention approach where query heads are partitioned into groups that share key-value heads, balancing efficiency between full multi-head attention…
neural architecture
GroupNorm is a normalization technique that divides channels into groups and normalizes within each group — independent of batch size, making it the preferred normalization…
quantum ai
Grover's Algorithm is a quantum search algorithm that finds a marked item in an unsorted database of N elements using only O(√N) queries to the database oracle, achieving…
grover search algorithm, grover search, search algorithm
Grover Search Algorithm Introduction & Motivation Grover Search Algorithm addresses a central problem in quantum process control and search-algorithm-quadratic-speedup prediction…
grover's search and optimization, grover's search optimization, grover's search, search optimization
Grover's Search and Optimization Introduction & Motivation Grover's Search and Optimization addresses a central problem in quantum-algorithms process control…
grown junction 1951 first n type region, quiet growth interval czochralski, emitter collector region length, grown junction transistor outer region growth
Grow the First N-Type Region: The First Quiet Interval in an Alternating Process Nothing chemically new happens during this step — the melt is still the same n-type composition…
grown junction 1951 second n type region, collector depletion width reverse bias, reach through breakdown transistor, grown junction transistor collector length
Grow the Second N-Type Region: A Quiet Interval With a Voltage-Dependent Constraint Step twelve's quiet interval had no electrical constraint at all; this one has a constraint…
grown junction 1951 thin p type base, base width current gain exponential, grown junction transistor gain sensitivity, diffusion length base width ratio
Grow the Thin P-Type Base: The Quiet Interval That Isn't Allowed to Drift Step twelve's quiet interval had no electrical constraint at all — this one has the tightest constraint…
grown-junction transistor process flow, grown junction transistor device process flow, grown junction transistor process, Czochralski melt dosed pull, Teal Sparks grown junction, pull rate base width control
Grown-Junction Transistor Device Process Flow: Seed-Dip Timing, Pull-Rate Kinetics & Dopant-Switch Yield Control The grown-junction transistor was the first transistor…
rpc, streaming
gRPC is the high-performance Remote Procedure Call framework developed by Google that uses HTTP/2 for transport and Protocol Buffers for serialization — enabling efficient…
group relative policy optimization, llm reward free rl, process reward model training, math reasoning rl
GRPO and RL for LLM Reasoning is the reinforcement learning training paradigm that directly optimizes large language models for verifiable reasoning tasks — particularly…
gru gated recurrent, gru gated, gated recurrent
GRU Gated Recurrent Introduction & Motivation GRU Gated Recurrent addresses a central problem in timeseries process control and gru-reset-update-gate prediction: how to Predict…
gru gated, gated recurrent, cnn gru lstm passed rnn, gru gated recurrent, gated recurrent unit, recurrent unit
GRU: Gated Recurrent Unit Introduction & Motivation GRU (Gated Recurrent Unit): simplified LSTM alternative. Reset gate: selective past information. Update gate: control…
recommendation systems
GRU4Rec is a session-based recommendation model using gated recurrent units over click sequences - Sequential hidden states encode short-term intent and predict next likely items…
evaluation
gSCAN (grounded SCAN) is the benchmark for systematically testing compositional generalization in visually grounded instruction following — placing an agent in a grid world where…
evaluation
GSM8K (Grade School Math 8K) is a benchmark of 8,500 high-quality, linguistically diverse grade school math word problems requiring 2-8 step reasoning to solve, designed…
math benchmark, word problems
GSM8K is a benchmark dataset of 8,500 linguistically diverse grade school math word problems designed to evaluate mathematical reasoning capabilities in language models. What Is…
gsm8k, evaluation
GSM8K is a grade-school math word-problem benchmark used to evaluate multi-step numerical reasoning - It is a core method in modern AI evaluation and safety execution workflows.…
Gradient-Based Sample Selection GSS for Memory Management Introduction & Motivation Gradient-Based Sample Selection GSS for Memory Management addresses a central problem…
gtn, graph neural networks
GTN is graph transformer network that learns soft meta-relational paths in heterogeneous graphs - It automates metapath construction instead of relying solely on hand-crafted…
qlora, efficient
Guanaco is a landmark family of fine-tuned language models that demonstrated QLoRA (Quantized Low-Rank Adaptation) could produce chatbot performance rivaling ChatGPT while…
ring, isolation, techniques, substrate, coupling
Guard Ring and Isolation Techniques is protective structures surrounding sensitive circuits reducing substrate and electromagnetic coupling — essential for noise-sensitive analog…
design
Guard band is a deliberate safety margin between nominal operation and failure limits — ensuring process drift, temperature swings, and aging don't push systems into unsafe…
guard band test limit optimization prediction, guard band test, band test limit, test limit optimization, limit optimization prediction, guard band, band test, test limit
Guard Band Test Limit Optimization Prediction with Machine Learning Introduction & Motivation Guard Band Test Limit Optimization Prediction with Machine Learning addresses…
design, latchup prevention, substrate resistance, minority carriers
CMOS latch-up constitutes the destructive, self-sustaining low-impedance state triggered by the regenerative turn-on of parasitic bipolar junction transistors inherent to bulk…
design, latch-up, substrate noise, well taps, minority carrier collection
CMOS latch-up constitutes the destructive, self-sustaining low-impedance state triggered by the regenerative turn-on of parasitic bipolar junction transistors inherent to bulk…
guardband reduction and yield-reliability trade-off, guardband reduction yield-reliability, reduction yield-reliability trade-off, guardband reduction, reduction yield-reliability
Guardband Reduction and Yield-Reliability Trade-Off Introduction & Motivation Guardband Reduction and Yield-Reliability Trade-Off addresses a central problem in design process…
design
Guardbanding is the intentional addition of design or test margin so products remain compliant under process, voltage, temperature, and aging uncertainty - it protects field…
advanced test & probe
Guardbanding is the practice of tightening test limits beyond nominal specifications to reduce defect escapes - It adds safety margin against measurement uncertainty, drift…
boundary, limit
Guardrails are the safety and compliance constraints that sit between users and language models to prevent harmful, off-topic, or policy-violating outputs — implemented as system…
ai safety
Guardrails is programmable constraints that enforce behavior, policy, and tool-usage limits in LLM workflows - It is a core method in modern AI safety execution workflows. What Is…
framework
Guardrails AI is the open-source framework for adding validation, safety checks, and structural constraints to LLM outputs — providing programmable guardrails that verify language…
structured, microsoft
Guidance is a Microsoft-developed programming language for constraining and controlling LLM outputs with guaranteed structure — replacing probabilistic prompt engineering…
framework
Guidance is the constraint-based language model programming framework by Microsoft that enables precise control over LLM output structure through interleaved generation…
Guidance-Free Diffusion Sampling Introduction & Motivation Guidance-Free Diffusion Sampling addresses a central problem in generative modeling of diffusion data: how to Learn…
Guidance-Free Diffusion Sampling (Extended Study) Introduction & Motivation Guidance-Free Diffusion Sampling (Extended Study) addresses a central problem in generative modeling…
Guidance-Free Diffusion Sampling (Extended Study) Introduction & Motivation Guidance-Free Diffusion Sampling (Extended Study) addresses a central problem in generative modeling…
generative models
Guidance scale is the numeric factor in classifier-free guidance that sets the strength of conditional steering during denoising - it is one of the most sensitive controls…
multimodal ai
Guidance Scale is the control parameter determining strength of conditional guidance during diffusion sampling - It directly affects prompt fidelity and output variability. What…
interpretability
Guided Backprop is a visualization method that modifies backpropagation to pass only positive gradients through ReLU layers - It produces sharper feature-importance maps than…