fisher-weighted averaging
model merging
Fisher-Weighted Averaging is a model merging technique that weights each parameter by its Fisher information — parameters that are more important for a task (higher Fisher…
1,257 technical terms and definitions
model merging
Fisher-Weighted Averaging is a model merging technique that weights each parameter by its Fisher information — parameters that are more important for a task (higher Fisher…
fism, recommendation systems
FISM is a factored item similarity model that predicts preferences from interactions with similar items - Item-item similarities are learned in latent space and aggregated…
fit, business & standards
Semiconductor reliability physics and accelerated life testing constitute the statistical, thermodynamic, and mechanical disciplines engineered to predict, quantify, and guarantee…
ai infrastructure stack, data center power economics, ai silicon and models, ai value capture layers, build vs buy ai
Five-Layer AI Market Stack describes how value is created from electricity to end-user applications, and why bottlenecks migrate across the stack over time. For 2024 to 2026…
quality & reliability
Five Whys is an iterative questioning method used to drill from an observed problem down to causal drivers - It supports fast structured investigation when data is limited. What…
reliability
Fix effectiveness factor is a measure of how strongly a corrective action reduces recurrence of a targeted failure mechanism - Effectiveness is estimated by comparing failure…
sparse attention
Fixed Attention Patterns are predetermined, static sparsity patterns for self-attention — where the set of positions each token can attend to is defined before training and does…
rag
Fixed-length chunking is the document splitting method that creates chunks by uniform token or character counts regardless of linguistic boundaries - it is simple and fast but can…
fpn imaging, photoresponse nonuniformity
Fixed pattern noise is the specific error correlated double sampling was built to fight — a spatial, repeatable deviation from ideal response across an array of nominally…
fixed-point arithmetic, q format, integer arithmetic, int8, int4, quantized inference
Fixed point arithmetic represents values as integers with an implicit scale, commonly a power-of-two binary point. It delivers predictable, compact, energy-efficient arithmetic…
reliability
Fixed-time life testing is life-testing approach where units are tested for a predetermined duration before reliability decisions are made - All samples run to the planned…
semi-supervised learning
FixMatch is a semi-supervised learning algorithm that combines pseudo-labeling with consistency regularization — using weak augmentation to generate confident pseudo-labels…
advanced training
FixMatch is a semi-supervised algorithm that combines weak-augmentation pseudo labels with strong-augmentation consistency training - High-confidence predictions from weakly…
FixMatch Weak-to-Strong Augmentation Consistency Introduction & Motivation FixMatch Weak-to-Strong Augmentation Consistency addresses a central problem in semisupervised…
FixMatch Semi-Supervised Algorithm Introduction & Motivation FixMatch Semi-Supervised Algorithm addresses a central problem in semisupervised and weak-strong-augmentation-design…
code ai
Fixture Generation is the AI task of automatically creating the test data setup and teardown code — database records, file contents, object instances, environment configurations…
optimization
Fixup Initialization is a weight initialization scheme for residual networks that enables stable training of arbitrarily deep networks without any normalization layers…
flaky test detection and isolation, flaky test detection, test detection isolation, flaky test, test detection, detection isolation
Flaky Test Detection and Isolation Introduction & Motivation Flaky Test Detection and Isolation addresses a central problem in verification process control…
cpu, visualization
Flame Graphs are the hierarchical visualization of software profiling data where bar width represents time spent (including children) and bar height represents call stack depth…
packaging
Flame retardant in EMC is the additive system in epoxy molding compounds that improves resistance to ignition and flame propagation - it helps packages meet safety and regulatory…
multimodal ai
Flamingo is a visual language model (VLM) developed by DeepMind — enabling few-shot learning for vision tasks by fusing a frozen pre-trained vision encoder and a frozen large…
flamingo multimodal in context learning, flamingo multimodal context learning, flamingo multimodal, multimodal context, coco llm qa vqa, flamingo multimodal in-context, multimodal in-context
Flamingo - Multimodal In-Context Learning Introduction & Motivation Flamingo: multimodal model with in-context learning. Visual and textual understanding. Applications: image-text…
flan, fine-tuning
FLAN is an instruction-tuning approach that mixes many prompted tasks to improve zero-shot and few-shot generalization - Task-diverse fine-tuning trains the model to follow…
flan, training techniques
FLAN is a fine-tuning paradigm that improves instruction following by training on diverse task instructions and formatted outputs - It is a core method in modern LLM training…
foundation model
FLAN-T5 is Google's instruction-tuned version of the T5 model, fine-tuned on a massive collection of diverse tasks described via natural language instructions, dramatically…
attention, efficient, transformer, mechanism
Flash Attention and Efficient Transformer Mechanisms is an optimized attention algorithm that reduces memory accesses and computation through IO-aware implementation — achieving…
flash adc metastability prediction, flash adc metastability, adc metastability prediction, flash adc, adc metastability, metastability prediction
Flash ADC Metastability Prediction with Machine Learning Introduction & Motivation Flash ADC Metastability Prediction with Machine Learning addresses a central problem in flash…
implant
Flash annealing uses millisecond-duration high-intensity light pulses to heat the wafer surface to extreme temperatures (1200-1350°C) while the bulk remains relatively cool…
process integration
Flash Anneal is an ultra-short high-temperature anneal using intense lamp pulses for rapid activation - It offers strong dopant activation while suppressing unwanted diffusion…
memory efficient
FlashAttention FlashAttention is a memory-efficient attention algorithm that fuses operations into optimized CUDA kernels reducing memory complexity from O of n squared to O of n…
memory efficient
FlashAttention is a breakthrough algorithm that computes exact attention without materializing the full N×N attention matrix, reducing memory from O(N²) to O(N) while achieving…
efficient attention, io-aware attention
FlashAttention is an IO-aware attention algorithm that computes exact attention while never writing the full N×N score matrix to HBM. Instead of materializing all pairwise scores…
memory efficient attention, io aware attention, flash attention v2
FlashAttention is a memory-efficient, IO-aware exact attention algorithm that reduces GPU memory usage from O(N²) to O(N) and speeds up attention computation by 2-4x — enabling…
io-aware attention, memory efficient attention, attention kernel optimization, fused attention cuda
Flash Attention is the IO-aware, exact attention algorithm that computes self-attention in O(N) memory (instead of O(N²)) by tiling the computation to exploit GPU SRAM (shared…
io aware attention, tiled attention, flash attention memory, fused attention kernel
Flash Attention is the IO-aware exact attention algorithm that computes multi-head self-attention without materializing the full N×N attention matrix in GPU HBM — using tiling…
memory efficient attention, attention kernel optimization, flash attention implementation, subquadratic attention
FlashAttention is an IO-aware attention algorithm that computes exact attention while never writing the full N×N score matrix to HBM. Instead of materializing all pairwise scores…
flash attention algorithm, memory efficient attention, tiling attention computation, flash attention gpu optimization
Flash Attention is the IO-aware exact attention algorithm that computes self-attention with O(N) memory instead of O(N²) and 2-4× faster wall-clock time than standard attention…
flash cell threshold distribution prediction, flash cell threshold, cell threshold distribution, threshold distribution prediction, flash cell, cell threshold
Flash Cell Threshold Distribution Prediction with Machine Learning Introduction & Motivation Flash Cell Threshold Distribution Prediction with Machine Learning addresses a central…
flash erase speed prediction, flash erase speed, erase speed prediction, flash erase, erase speed, speed prediction
Flash Erase Speed Prediction with Machine Learning Introduction & Motivation Flash Erase Speed Prediction with Machine Learning addresses a central problem in flash memory erase…
flash lamp anneal dwell time optimization prediction, flash lamp anneal, lamp anneal dwell, anneal dwell time, dwell time optimization, flash lamp, lamp anneal
Flash Lamp Anneal Dwell Time Optimization Prediction with Machine Learning Introduction & Motivation Flash Lamp Anneal Dwell Time Optimization Prediction with Machine Learning…
NAND flash, NOR flash, Fujio Masuoka, floating-gate transistor, erase block, wear leveling, 3D NAND
Flash memory is a non-volatile storage technology that stores each bit, or several bits, as electric charge trapped on an electrically isolated gate inside a transistor, keeping…
flash memory bit error rate prediction, flash memory bit, memory bit error, bit error rate, error rate prediction, flash memory, memory bit, bit error, error rate
Flash Memory Bit Error Rate Prediction with Machine Learning Introduction & Motivation Flash Memory Bit Error Rate Prediction with Machine Learning addresses a central problem…
floating gate transistor, charge trap flash, ctf, sonos flash, nand flash cell
Flash Memory Cell Process is the fabrication sequence for nonvolatile storage transistors that trap charge either in a polysilicon floating gate or in a nitride charge-trap layer…
flash memory endurance cycling degradation prediction machine learning, flash memory, endurance cycling degradation, nand, flash memory endurance, memory endurance
Flash Memory Endurance Cycling Degradation Prediction with Machine Learning Introduction & Motivation Flash Memory Endurance Cycling Degradation Prediction with Machine Learning…
flash memory wear leveling effectiveness prediction, flash memory wear, memory wear leveling, wear leveling effectiveness, leveling effectiveness prediction
Flash Memory Wear Leveling Effectiveness Prediction with Machine Learning Introduction & Motivation Flash Memory Wear Leveling Effectiveness Prediction with Machine Learning…
flash program speed prediction, flash program speed, program speed prediction, flash program, program speed, speed prediction
Flash Program Speed Prediction with Machine Learning Introduction & Motivation Flash Program Speed Prediction with Machine Learning addresses a central problem in flash memory…
flash retention endurance prediction, flash retention endurance, retention endurance prediction, flash retention, retention endurance, endurance prediction
Flash Retention Endurance Prediction with Machine Learning Introduction & Motivation Flash Retention Endurance Prediction with Machine Learning addresses a central problem…
optimization
FlashAttention-2 is an optimized implementation of the attention mechanism that achieves 2× the speed of the original FlashAttention — reaching up to 230 TFLOPS/s on NVIDIA A100…
optimization
FlashAttention implementation is the IO-aware exact attention algorithm design that computes attention in tiled blocks to avoid full score-matrix storage - it delivers large speed…
memory, spaced
AI flashcard generation and spaced repetition automatically converts content into memorization cards — using AI to create high-quality flashcards from notes, PDFs, or videos, then…