← Back to AI Factory Chat

AI Factory Glossary

422 technical terms and definitions

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z All
Showing page 2 of 9 (422 entries)

edge pooling, graph neural networks

Edge pooling contracts edges to coarsen graphs learning which edges to collapse for hierarchical representations.

edge pooling, graph neural networks

Pool edges to coarsen graph.

edge popup,model optimization

Discover subnetworks via masking.

edge probing, interpretability

Edge probing evaluates whether representations capture syntactic relationships between word spans.

edge rounding, process

Smooth wafer edge.

edge trim, process

Remove wafer edge.

edge-cloud collaboration, edge ai

Split computation between edge and cloud.

edge,on device,local inference

Edge inference runs models on device (phone, IoT). Lower latency, offline capable, privacy. Requires smaller models.

edi, edi, supply chain & logistics

Electronic Data Interchange automates business document exchange between supply chain partners.

edit-based generation, text generation

Generate by editing initial sequence.

editing models via task vectors, model merging

Modify capabilities using task directions.

editing real images with gans, generative models

Modify real images via inversion.

edt, edt, design & verification

Embedded Deterministic Test generates pseudo-random patterns with controlled bits.

eeg analysis,healthcare ai

Analyze brain wave patterns.

eend, eend, audio & speech

End-to-End Neural Diarization performs joint speech activity and speaker detection using single model.

efem (equipment front end module),efem,equipment front end module,automation

Mini-cleanroom at tool front with robot for wafer handling.

efem, efem, manufacturing operations

Equipment Front End Module provides automated wafer handling interface.

effect size, quality & reliability

Effect sizes quantify practical significance of differences or relationships.

effective mass calculation, simulation

Determine carrier effective masses.

effective potential method, simulation

Another quantum correction approach.

efficient attention mechanisms for vit, computer vision

Various efficient attention variants.

efficient attention variants,llm architecture

Attention mechanisms faster than O(n²).

efficient net,mobile,edge

EfficientNet family balances accuracy and efficiency. Compound scaling. Good for mobile/edge.

efficient neural architecture search, enas, neural architecture

Share weights across architectures during search.

efficientnet nas, neural architecture search

EfficientNet architecture discovered through NAS balances depth width and resolution using compound scaling.

efficientnet scaling, model optimization

EfficientNet scaling jointly optimizes depth width and resolution using compound coefficient.

efficientnet, computer vision

Scale networks systematically.

efficientnetv2, computer vision

Improved training and scaling.

egnn, egnn, graph neural networks

E(n) Equivariant Graph Neural Networks maintain equivariance to Euclidean transformations while learning node and edge features in geometric graphs.

eigen-cam, explainable ai

Use principal components for CAM.

eight points beyond zone c, spc

Excessive variation.

einstein relation, device physics

Relation between mobility and diffusion.

einstein, e=mc2, mass energy equivalence, relativity, physics

# Einstein's Mass-Energy Equivalence ## E = mc² Einstein's mass–energy equivalence states that mass and energy are interconvertible according to: $$ E = mc^2 $$ Where: - $E$ = energy (joules, J) - $m$ = mass (kilograms, kg) - $c$ = speed of light in vacuum ≈ $3.00 \times 10^8$ m/s ## What It Means Mass and energy are two manifestations of the same fundamental quantity. A body with mass $m$ possesses an intrinsic "rest energy" even in its stationary state: $$ E_0 = mc^2 $$ Since $c^2$ is enormous (~$9 \times 10^{16}$ m²/s²), a minuscule amount of mass corresponds to an enormous quantity of energy. ## Common Misconception The equation does **not** state that "mass converts to energy only at high velocities." The formula describes the rest mass energy inherent to all matter. When objects move, the total energy becomes larger; in relativistic mechanics: $$ E^2 = (pc)^2 + (mc^2)^2 $$ where $p$ is relativistic momentum. ## Practical Examples - **1 kg of mass** = $9 \times 10^{16}$ joules (equivalent to ~21 megatons of TNT) - **Nuclear fission**: Converts ~0.1% of mass to energy - **Nuclear fusion**: Converts ~0.7% of mass to energy - **Matter-antimatter annihilation**: Converts 100% of mass to energy (complete conversion) ## Semiconductor Physics Context Mass-energy equivalence appears in semiconductor physics and device engineering: - **Electron rest mass energy**: $m_0 c^2 \approx 0.511$ MeV - **Relativistic band structure**: Heavy elements require relativistic corrections for accurate electronic structure - **Pair production thresholds**: Radiation damage and high-energy particle interactions - **Positron emission tomography (PET)**: Advanced defect imaging and characterization techniques - **Ion implantation**: Energetic particle transport relies on relativistic dynamics ## Historical Significance Einstein's equation fundamentally transformed our understanding of matter, energy, and the universe. It enabled breakthrough technologies including: - Nuclear power generation - Particle accelerators (linear accelerators for ion implantation) - Medical imaging (PET, radiation therapy) - Astrophysics and stellar processes

eisner algorithm, structured prediction

Eisner algorithm efficiently finds the highest-scoring projective dependency tree using dynamic programming over chart structures.

elastic modulus prediction, materials science

Predict mechanical properties.

elastic net attack, ai safety

L1+L2 regularized attack.

elastic recoil detection (erd),elastic recoil detection,erd,metrology

Measure hydrogen and light elements.

elastic weight consolidation (ewc),elastic weight consolidation,ewc,model training

Regularization technique to prevent catastrophic forgetting by protecting important weights.

elastic weight consolidation, ewc, continual learning

Protect important weights.

elbow method, manufacturing operations

Elbow method identifies optimal cluster count by plotting within-cluster variance.

electra generator-discriminator, electra, foundation model

Train discriminator to detect replaced tokens.

electra,foundation model

Efficient pre-training using discriminative task instead of masking.

electrical test structures,metrology

On-wafer structures for measuring parameters.

electrical wafer sort (ews),electrical wafer sort,ews,testing

Test dies electrically on wafer.

electrical width, yield enhancement

Electrical width is measured linewidth from resistance or capacitance differing from physical dimension.

electrochemical migration, reliability

Metal transport under bias.

electrochemical plating (ecp),electrochemical plating,ecp,beol

Deposit copper by electroplating.

electrodeionization, environmental & sustainability

Electrodeionization combines ion exchange and electrodialysis for continuous ultra-pure water production.

electroless plating,beol

Chemical plating without external current.

electroluminescence, el, metrology

Light emission from electrical injection.