experiment
iterate, feedback loop
Experimentation and Iteration The Build-Measure-Learn Loop For AI Applications Types of Experiments Prompt Experiments Test different system prompts Compare few-shot examples Try…
29,259 technical terms and definitions
iterate, feedback loop
Experimentation and Iteration The Build-Measure-Learn Loop For AI Applications Types of Experiments Prompt Experiments Test different system prompts Compare few-shot examples Try…
mlops
Experiment configuration management is the discipline of defining, versioning, validating, and governing all settings that determine experiment behavior - it prevents…
ml experiment tracking, weights and biases, wandb, mlflow, neptune, comet, tensorboard, clearml
Experiment tracking records the configuration, metrics, artifacts and provenance of model-development runs so results can be compared and reproduced. Without tracking, teams…
experiment tracking metadata, experiment tracking, tracking metadata
Experiment Tracking Metadata Introduction & Motivation Experiment Tracking Metadata addresses a central problem in ml-ops process control and tracking-overhead-performance…
factorial, randomization, block design, orthogonal array
Experimental design is the disciplined science of planning experiments so that the data they produce answer the engineer's question with maximum efficiency and minimum bias…
experimental mos bipolar logic step 1 why resistor transistor logic draws current in both states, rtl standing power baseline
Confront the 1962 Bipolar Logic Baseline: Why a Field-Effect Alternative Looked Worth the Risk 1. Why 1962's Digital Logic Was Entirely Bipolar, and Entirely Hungry for…
experimental mos bipolar logic step 2 why a workable field effect device took three decades to build, lilienfeld field effect patent history
Revisit the Field-Effect Concept Dormant Since the 1930s Patents 1. Why a Workable Device Took Three Decades Longer Than the Idea Behind It Julius Edgar Lilienfeld's 1926…
experimental mos bipolar logic step 3 surface passivated mos structure first working device, atalla kahng mos transistor 1960
Build the First Working MOS Transistor: Atalla and Kahngs Surface-Passivated Structure 1. Why This Specific Structure, Built at Bell Labs in 1959-1960, Finally Worked Mohamed…
experimental mos bipolar logic step 4 why early mos devices had unpredictable threshold voltage, mos threshold voltage surface state charge
Characterize Threshold Voltage and Residual Surface-State Charge in the Early Oxide 1. Why the First Devices Worked but Did Not Behave Predictably Even with a passivated surface…
experimental mos bipolar logic step 5 positive oxide charge forced early pmos polarity choice, early pmos channel polarity choice
Confront Early PMOS Polarity: Why the First Working Devices Were P-Channel, Not N-Channel 1. Why Every Early Experimental MOS Device Used a P-Channel, Against the Intuitive Choice…
experimental mos bipolar logic step 6 dynamic charge storage shift register first mos ic demonstration, rca experimental pmos shift register
Lay Out RCAs Experimental PMOS Shift Register: The First MOS Integrated Logic Demonstration 1. Why a Shift Register, Specifically, Was the First Circuit Built This Way RCA's…
experimental mos bipolar logic step 7 half cv squared switching energy versus rtl standing power, mos switching energy versus rtl power
Compare MOS Gate-Capacitance Switching Energy Against Bipolar RTL Switching Energy 1. Why the Honest Energy Comparison Is Not as Favorable to MOS as It Later Became Switching…
experimental mos bipolar logic step 8 rtl nor gate topology and fan out noise margin cost, resistor transistor logic rtl baseline
Define Resistor-Transistor Logic: The Incumbent Bipolar Baseline This Series Compares Against 1. Why RTL's Specific Circuit Topology Matters, Not Just Its Name Resistor-transistor…
experimental mos bipolar logic step 9 measured rtl noise margin falling with fan out count, rtl fan out noise margin limit
Confront RTLs Fan-Out and Noise-Margin Limits 1. Why RTL's Noise Margin Was the Family's Actual Practical Ceiling Measured across real RTL gates driving real fan-out loads…
experimental mos bipolar logic step 10 diode input clamp fixes fan out dependent noise margin, diode transistor logic dtl noise margin
Define Diode-Transistor Logic: RTLs Direct Answer to Its Own Noise-Margin Problem 1. Why Swapping Resistors for Diodes at the Input Specifically Fixes the Measured Problem…
experimental mos bipolar logic step 11 dtl delay cost versus early pmos gate delay and power, dtl delay versus early pmos gate delay
Compare DTL Propagation Delay and Power Against Early PMOS Gate Delay 1. Why DTL's Noise-Margin Fix Added Delay That Early PMOS Did Not Pay DTL's diode input network adds its own…
experimental mos bipolar logic step 12 fairchild static mos logic gate independent confirmation, fairchild experimental mos logic gate
Fabricate Fairchilds Experimental MOS Logic Gate: A Second Independent Confirmation 1. Why a Second Lab Reaching the Same Result Mattered More Than the First Fairchild…
experimental mos bipolar logic step 13 mobile sodium ion migration under sustained gate bias, sodium ion mos threshold instability
Confront Early MOS Threshold Instability: Sodium-Ion Contamination and Oxide Drift 1. Why Devices That Worked on the Bench Failed Under Sustained Bias Devices that measured…
experimental mos bipolar logic step 14 phosphosilicate glass sodium getter layer fix, phosphosilicate glass psg passivation fix
Solve Threshold Drift With Phosphosilicate Glass Passivation 1. Why a Phosphorus-Doped Glass Layer Specifically Stops Sodium Migration Once sodium contamination was identified as…
experimental mos bipolar logic step 15 fewer component types per mos nor gate than bipolar, mos gate density versus bipolar density
Scale Experimental MOS Gate Density Against Bipolar RTL and DTL Gate Density 1. Why a Single MOS Transistor Could Replace Several Bipolar Components Per Gate An RTL or DTL NOR…
experimental mos bipolar logic step 16 mos off state leakage versus bipolar standing current, mos static power versus bipolar standby power
Compare Experimental MOS Static Power Against Bipolar Standby Power 1. Why Static Power, Not Switching Power, Was MOS's Clearest Measured Advantage Measuring a PMOS gate and…
experimental mos bipolar logic step 17 low mobility thick oxide and channel resistance slow early pmos, early pmos speed deficit versus bipolar
Confront the Experimental MOS Speed Deficit Against Bipolar Switching Speed 1. Why Early PMOS Could Not Yet Match Bipolar Switching Speed at Any Reasonable Voltage Measured…
experimental mos bipolar logic step 18 threshold scatter contamination and oxide pinhole defects limit yield, experimental mos yield production limit
Characterize Yield: Why Experimental MOS Struggled to Reach Production Volume 1. Why Laboratory Success Did Not Translate Directly Into a Shippable Product Line Across RCA's…
experimental mos bipolar logic step 19 four independent labs divided the experimental mos problem, competing mos development programs 1960s
Survey the Competing 1962-1965 MOS Development Programs: RCA, Fairchild, GE, and Bell Labs 1. Why Multiple Independent Programs, Not One, Carried This Era Forward Beyond RCA's…
experimental mos bipolar logic step 20 honest split scoreboard mos density power versus bipolar speed yield, experimental mos bipolar logic series close
Close on Experimental MOS's 1962-1965 Arc: A Real Capability, Not Yet a Winning Product 1. Why This Series Ends on an Honest, Unresolved Competitive Standing, Not a Verdict By…
data
Expert annotation is the process of having domain specialists — such as doctors, lawyers, linguists, or engineers — create labeled training and evaluation data for machine…
moe
Expert Capacity is the maximum number of tokens that can be routed to any single expert within a Mixture-of-Experts (MoE) layer during a single forward pass, defined as…
architecture
Expert Capacity is maximum token budget assigned to each expert within a sparse mixture layer - It is a core method in modern semiconductor AI serving and inference-optimization…
moe
Expert Capacity Factor is the hyperparameter in Mixture of Experts (MoE) models that controls the maximum number of tokens each expert can process per batch — calculated as (total…
moe
Expert Choice Routing is the MoE routing paradigm that inverts the traditional token-selects-expert direction — instead, each expert independently selects the top-k tokens it…
moe
Expert dropout is the regularization technique that temporarily disables a subset of experts during training to reduce over-reliance on dominant experts - it encourages more…
moe
Expert load balancing is the process of distributing routed tokens across experts so no small subset becomes overloaded while others idle - it is essential for achieving both…
moe, switch
Expert parallelism distributes Mixture-of-Experts (MoE) model experts across different GPUs, enabling sparse activation with massive total parameter counts. Architecture: router…
distributed training
Expert parallelism is a distributed computing strategy specifically designed for Mixture of Experts (MoE) models, where different expert sub-networks are placed on different GPUs.…
architecture
Expert Parallelism is distributed execution strategy that shards experts across multiple devices for scalable sparse training - It is a core method in modern semiconductor AI…
moe
Expert parallelism implementation is the distributed execution strategy that shards experts across devices while sharing router work across replicas - it allows sparse models…
mixture experts parallelism, moe distributed training, expert placement strategies, load balancing experts
Expert parallelism distributes an MoE model's many separate expert sub-networks across multiple accelerator devices, with each device hosting and computing only a specific subset…
mixture of experts distributed, moe training parallelism, expert model parallel, switch transformer training
Expert parallelism's efficiency depends heavily on how evenly a router actually distributes tokens across the available experts, since an imbalanced routing pattern that sends…
moe
Expert redundancy is the undesired condition where multiple MoE experts learn highly overlapping functions, reducing effective sparse capacity - it limits quality gains and wastes…
model architecture
Expert routing determines which experts process each token in Mixture of Experts architectures. Router network: Small network (often single linear layer) that takes token…
architecture
Expert Routing is process of assigning each token to one or more specialized experts in sparse architectures - It is a core method in modern semiconductor AI serving…
moe
Expert specialization is the emergent behavior where different MoE experts learn distinct token or task sub-distributions over training - it is the main mechanism by which sparse…
knowledge-based system, MYCIN, XCON, R1, DENDRAL, inference engine, knowledge acquisition bottleneck, CLIPS, EMYCIN
An expert system is a deployed knowledge-based program that captures the specialized judgment of human experts in a narrow domain, typically as a knowledge base of if-then rules…
Large Language Model Training Modern LLM training follows a systematic approach from data to deployment: Training Pipeline Overview Large Language Model training is a multi-stage…
interpretability vs explainability, mechanistic interpretability
Explainability is a model's ability to have its decisions understood by a human — not just "it predicted X," but why it predicted X, in terms a person can actually reason about…
explainability interpretability shap lime attention saliency, explainability interpretability, shap lime attention, 10k ai bert cam clip compas, lime attention saliency
SHAP, LIME, attention visualization, and saliency maps represent different approaches to post-hoc explainability, each attributing a model's output to specific input features…
Explainability Evaluation Metrics Introduction & Motivation Explainability Evaluation Metrics addresses a central problem in interpretability and explanation of xai predictions…
Explainability Evaluation Metrics (Extended Study) Introduction & Motivation Explainability Evaluation Metrics (Extended Study) addresses a central problem in interpretability…
Explainability Evaluation Metrics (Extended Study) Introduction & Motivation Explainability Evaluation Metrics (Extended Study) addresses a central problem in interpretability…
Explainability for Graph Neural Networks Introduction & Motivation Explainability for Graph Neural Networks addresses a central problem in interpretability and explanation of xai…