F
101 words and 273 other names, in italics, pointing to their entries.
- F1 → F1 score
- F1 score
- fab → semiconductor foundry
- fabless → semiconductor foundry
- fabric → interconnect
- fabrication → confabulation
- FAccT
- face detection
- face detector → face detection
- face identification → facial recognition
- face recognition → facial recognition
- face swap → deepfake
- face verification → facial recognition
- Facebook AI Research → Meta AI
- Facebook AI Similarity Search → FAISS
- faceprint → facial recognition
- facial landmark detection → face detection
- facial recognition
- facial recognition technology → facial recognition
- fact editing → model editing
- factor graph → probabilistic graphical model
- factored cognition
- factual accuracy → factuality
- factual consistency → factuality
- factual recall → recall
- factuality
- factuality evaluation → factuality
- FAI → friendly AI
- fail-slow → straggler
- failure-time analysis → survival analysis
- FAIR → Meta AI
- fair machine learning → algorithmic fairness
- fair use and AI training
- fair use defense for AI training → fair use and AI training
- fairness impossibility results → fairness impossibility theorem
- fairness impossibility theorem
- fairness in machine learning → algorithmic fairness
- fairness metric → algorithmic fairness
- fairness through awareness → individual fairness
- fairness through blindness → fairness through unawareness
- fairness through unawareness
- fairness-aware machine learning → algorithmic fairness
- FAISS
- faithful → faithfulness
- faithfulness
- fake alignment → alignment faking
- fake Drake song → Heart on My Sleeve
- fake quantization → quantization-aware training
- faking alignment → alignment faking
- fall-out → false positive
- false alarm → false positive
- false discovery rate → false positive
- false negative
- false negative rate → false negative
- false positive
- false positive rate → false positive
- false refusal → over-refusal
- false wake → wake word
- false-belief task → theory of mind
- fancy autocomplete
- Fashion-MNIST → MNIST
- fast and slow thinking → System 1 and System 2
- fast Fourier transform → Fourier transform
- fast gradient sign attack → fast gradient sign method
- fast gradient sign method
- Fast R-CNN → R-CNN
- fast takeoff
- fast weights → hypernetwork
- Faster R-CNN → R-CNN
- faster-whisper → Whisper
- FastICA → independent component analysis
- FAT* → FAccT
- FAT/ML → FAccT
- F-beta score → F1 score
- Fβ score → F1 score
- FC layer → fully connected layer
- FCN → fully convolutional network
- FD-DINOv2 → Fréchet inception distance
- FDE → forward deployed engineer
- FDT → decision theory
- feature
- feature absorption
- feature attribution
- feature binding → binding problem
- feature circuit → sparse feature circuit
- feature clamping → feature steering
- feature construction → feature engineering
- feature creation → feature engineering
- feature drift → data drift
- feature engineering
- feature extraction
- feature extractor → feature extraction
- feature importance
- feature (interpretability)
- feature learning → representation learning
- feature map
- feature platform → feature store
- feature projection → dimensionality reduction
- feature reconstruction loss → perceptual loss
- feature scaling
- feature selection
- feature split → feature splitting
- feature splitting
- feature steering
- feature store
- feature superposition → superposition
- feature transformation → feature engineering
- feature vector → feature
- feature visualization
- feature viz → feature visualization
- features → feature
- federal preemption of state AI laws → AI preemption
- federated learning
- federated training → federated learning
- feedback control → control theory
- feedback loop
- feedback tampering → reward tampering
- feed-forward layer → feed-forward network
- feed-forward network
- feedforward neural network
- feed-forward neural network → feedforward neural network
- feel the AGI
- feeling the AGI → feel the AGI
- Feigenbaum bottleneck → knowledge acquisition bottleneck
- Edward Feigenbaum
- FEP → free energy principle
- feudal reinforcement learning → hierarchical reinforcement learning
- few-shot → few-shot prompting
- few-shot learning
- few-shot prompting
- few-step diffusion → diffusion distillation
- FFN → feed-forward network
- FFNN → feedforward neural network
- FGCS → Fifth Generation Computer Systems project
- FGSM → fast gradient sign method
- FHE → homomorphic encryption
- FHI → Future of Humanity Institute
- fictitious self-play → self-play
- FID → Fréchet inception distance
- FID score → Fréchet inception distance
- fidelity → faithfulness
- field-programmable gate array → FPGA
- Fifth Generation Computer Systems project
- Fifth Generation project → Fifth Generation Computer Systems project
- fifth-generation computer → Fifth Generation Computer Systems project
- fill-in-the-middle
- filter
- filter bank → filter
- filter bubble
- filterbank features → mel spectrogram
- final goal → terminal goal
- final run → training run
- Finch → RWKV
- fine-grained structured sparsity → N:M sparsity
- fine-tune → fine-tuning
- fine-tuned → fine-tuning
- fine-tuning
- finetuning → fine-tuning
- fine-tuning attack
- FineWeb
- FineWeb2 → FineWeb
- FineWeb-Edu → FineWeb
- finite mixture model → Gaussian mixture model
- Chelsea Finn
- fire alarm → warning shot
- first incompleteness theorem → Gödel's incompleteness theorems
- first-order approximation → Taylor series
- first-order logic
- first-order MAML → model-agnostic meta-learning
- first-order predicate calculus → first-order logic
- first-order predicate logic → first-order logic
- Fisher discriminant analysis → linear discriminant analysis
- Fisher information
- Fisher information matrix → Fisher information
- Fisher's linear discriminant → linear discriminant analysis
- fitness → fitness function
- fitness function
- fitness landscape → fitness function
- fitness score → fitness function
- FK → forward kinematics
- Flan-T5 → T5
- Flash Attention → FlashAttention
- flash attention → FlashAttention
- Flash Fill → program synthesis
- flash lidar → lidar
- FlashAttention
- FlashAttention-2 → FlashAttention
- FlashAttention-3 → FlashAttention
- flat minima → sharp and flat minima
- flat minimum → sharp and flat minima
- flatness → sharp and flat minima
- Flax → JAX
- FLI → Future of Life Institute
- FLI open letter → Pause Giant AI Experiments
- float → FP32
- float16 → FP16
- float32 → FP32
- float8 → FP8
- floating point → floating-point format
- floating-point format
- floating-point number → floating-point format
- floating-point operation → FLOP
- floating-point operations per second → FLOPS
- FLOP
- FLOP budget → compute budget
- FLOP per byte → arithmetic intensity
- FLOP threshold → compute threshold
- FLOPS
- FLOP/s → FLOPS
- FLOPs utilization → model FLOPs utilization
- flow field → optical flow
- flow matching
- flow model → normalizing flow
- flow-based generative model → normalizing flow
- flow-matching model → flow matching
- fluent → situation calculus
- fluid intelligence → g factor
- FLUX
- FLUX 3 → FLUX
- FLUX.1 Kontext → FLUX
- FM → foundation model
- FMCW lidar → lidar
- F-measure → F1 score
- FMF → Frontier Model Forum
- FMOps → LLMOps
- FNN → feedforward neural network
- FNO → neural operator
- FNR → false negative
- focal loss
- focal loss function → focal loss
- Jerry Fodor
- FOIL → inductive logic programming
- FOL → first-order logic
- folk psychology
- FOMAML → model-agnostic meta-learning
- FOOM
- force control → impedance control
- forecasting → time series forecasting
- Forethought
- Forethought Research → Forethought
- forgetting → catastrophic forgetting
- formal methods → formal verification
- formal verification
- formalization → interactive theorem proving
- formalized mathematics → interactive theorem proving
- format sensitivity → prompt sensitivity
- forward algorithm → hidden Markov model
- forward chaining
- forward computation → forward pass
- forward deployed engineer
- forward deployed engineering → forward deployed engineer
- forward kinematics
- forward KL → Kullback–Leibler divergence
- forward pass
- forward propagation → forward pass
- forward reasoning → forward chaining
- forward–backward algorithm → hidden Markov model
- forward-chain → forward chaining
- forward-mode differentiation → automatic differentiation
- foundation model
- foundation model for time series → time series foundation model
- Foundation Models framework → Core ML
- foundational ontology → ontology
- foundry → semiconductor foundry
- four-fifths rule → disparate impact
- Fourier analysis → Fourier transform
- Fourier neural operator → neural operator
- Fourier series → Fourier transform
- Fourier transform
- FP16
- FP32
- FP4
- FP64 → floating-point format
- FP8
- FP8 training → mixed-precision training
- FPGA
- FP-growth → association rule learning
- FPR → false positive
- fprop → forward pass
- FPV drone → drone
- fractionally strided convolution → transposed convolution
- frame
- frame axioms → frame problem
- frame problem
- frame system → frame
- frames → frame
- Framework Convention on Artificial Intelligence
- Framework for Artificial Intelligence Diffusion → AI Diffusion Rule
- Frankenstein complex
- Frankenstein syndrome → Frankenstein complex
- Fréchet audio distance → Fréchet inception distance
- Fréchet inception distance
- Fréchet video distance → Fréchet inception distance
- Fred Jelinek → Frederick Jelinek
- free energy principle
- free space → configuration space
- frequency domain → Fourier transform
- frequency penalty → repetition penalty
- frequent itemset mining → association rule learning
- FRIA → fundamental rights impact assessment
- Jerome Friedman
- friendliness → friendly AI
- friendly AI
- Karl Friston
- Frobenius norm → norm
- front-door criterion → do-calculus
- frontier AI
- frontier AI company → frontier lab
- frontier AI developer → frontier lab
- frontier AI lab → frontier lab
- frontier AI model → frontier model
- Frontier AI Safety Commitments
- frontier AI safety policy → responsible scaling policy
- frontier AI system → frontier AI
- frontier artificial intelligence → frontier AI
- frontier developer → frontier lab
- frontier lab
- frontier LLM → frontier model
- frontier model
- Frontier Model Forum
- frontier safety evaluation → dangerous capability evaluation
- Frontier Safety Framework
- frontier safety policy → responsible scaling policy
- FrontierMath
- FrontierMath: Open Problems → FrontierMath
- FrontierMath Tier 4 → FrontierMath
- frozen features → feature extraction
- FRT → facial recognition
- F-score → F1 score
- FSD → Tesla
- FSDP → fully sharded data parallel
- FSDP2 → fully sharded data parallel
- FSF → Frontier Safety Framework
- FSL → few-shot learning
- FTU → fairness through unawareness
- Kunihiko Fukushima
- full fine-tuning
- full finetuning → full fine-tuning
- full rank → rank
- Full Self-Driving → Tesla
- full-batch gradient descent → gradient descent
- full-duplex speech model → speech-to-speech model
- full-parameter fine-tuning → full fine-tuning
- full-weight fine-tuning → full fine-tuning
- fully connected layer
- fully connected network → multilayer perceptron
- fully convolutional network
- fully convolutional neural network → fully convolutional network
- fully homomorphic encryption → homomorphic encryption
- fully sharded data parallel
- FullyShardedDataParallel → fully sharded data parallel
- function call → tool call
- function calling
- function class → hypothesis space
- function of a function rule → chain rule
- function vector
- function vectors → function vector
- functional decision theory → decision theory
- functional margin → margin
- functionalism
- functionalist → functionalism
- Fundamental AI Research → Meta AI
- fundamental rights impact assessment
- Funk SVD → matrix factorization
- FunSearch
- fused kernel → kernel fusion
- Future of Humanity Institute
- Future of Life Institute
- fuzzy control → fuzzy logic
- fuzzy logic
- fuzzy logic controller → fuzzy logic
- fuzzy set → fuzzy logic
- fuzzy set theory → fuzzy logic
- FV → function vector