deduplication.tech

/

a HackerNoon project

BUG Insect ▸ Software bug MOUSE Mouse (animal) ▸ Computer mouse TOKEN Token (coin) ▸ Token (language model)BIAS Bias (prejudice) ▸ Bias (machine learning)TRANSFORMER Electrical transformer ▸ Transformer (deep learning)AGENT Agent (person) ▸ AI agent GROUNDING Grounding (punishment or electrical) ▸ Grounding (AI)EPOCH Epoch (era) ▸ Epoch (machine learning)INFERENCE Inference (logic) ▸ Inference (AI)DAEMON Daemon (mythology) ▸ Daemon (computing)MEMORY Memory (mind) ▸ Computer memory CACHE Cache (hidden store) ▸ Cache (computing)VIRUS Virus (biology) ▸ Computer virus PATCH Patch (fabric) ▸ Patch (computing)PING Ping (sound or sonar) ▸ Ping (networking)BRANCH Branch (tree) ▸ Branch (version control)PUSH Push (physical) ▸ Push (version control)SOCKET Socket (outlet or anatomy) ▸ Network socket TERMINAL Terminal (transport) ▸ Computer terminal KEY Key (lock) ▸ Key (cryptography)SEED Seed (botany) ▸ Random seed CLIENT Client (customer) ▸ Client (computing)FARM Farm ▸ Server farm TROLL Troll (folklore) ▸ Internet troll PHISH Fishing ▸ Phishing CRAWL Crawl (movement) ▸ Web crawl NET Net (mesh) ▸ The Net (internet)FOOTPRINT Footprint ▸ Memory footprint LEAK Leak (fluid) ▸ Memory leak EXCEPTION Exception (special case) ▸ Exception (programming)TREE Tree ▸ Tree (data structure)FEED Feed (animal) ▸ News feed HANDLE Handle (grip) ▸ Handle (username)ZOMBIE Zombie (folklore) ▸ Zombie (computing)CANARY Canary (bird) ▸ Canary release CLASS Class (school or society) ▸ Class (programming)CORE Core (center) ▸ Core (CPU)DECAY Decay (rot) ▸ Weight decay DOCK Dock (maritime) ▸ Dock (macOS)DRAWER Drawer (furniture) ▸ Navigation drawer BUG Insect ▸ Software bug MOUSE Mouse (animal) ▸ Computer mouse TOKEN Token (coin) ▸ Token (language model)BIAS Bias (prejudice) ▸ Bias (machine learning)TRANSFORMER Electrical transformer ▸ Transformer (deep learning)AGENT Agent (person) ▸ AI agent GROUNDING Grounding (punishment or electrical) ▸ Grounding (AI)EPOCH Epoch (era) ▸ Epoch (machine learning)INFERENCE Inference (logic) ▸ Inference (AI)DAEMON Daemon (mythology) ▸ Daemon (computing)MEMORY Memory (mind) ▸ Computer memory CACHE Cache (hidden store) ▸ Cache (computing)VIRUS Virus (biology) ▸ Computer virus PATCH Patch (fabric) ▸ Patch (computing)PING Ping (sound or sonar) ▸ Ping (networking)BRANCH Branch (tree) ▸ Branch (version control)PUSH Push (physical) ▸ Push (version control)SOCKET Socket (outlet or anatomy) ▸ Network socket TERMINAL Terminal (transport) ▸ Computer terminal KEY Key (lock) ▸ Key (cryptography)SEED Seed (botany) ▸ Random seed CLIENT Client (customer) ▸ Client (computing)FARM Farm ▸ Server farm TROLL Troll (folklore) ▸ Internet troll PHISH Fishing ▸ Phishing CRAWL Crawl (movement) ▸ Web crawl NET Net (mesh) ▸ The Net (internet)FOOTPRINT Footprint ▸ Memory footprint LEAK Leak (fluid) ▸ Memory leak EXCEPTION Exception (special case) ▸ Exception (programming)TREE Tree ▸ Tree (data structure)FEED Feed (animal) ▸ News feed HANDLE Handle (grip) ▸ Handle (username)ZOMBIE Zombie (folklore) ▸ Zombie (computing)CANARY Canary (bird) ▸ Canary release CLASS Class (school or society) ▸ Class (programming)CORE Core (center) ▸ Core (CPU)DECAY Decay (rot) ▸ Weight decay DOCK Dock (maritime) ▸ Dock (macOS)DRAWER Drawer (furniture) ▸ Navigation drawer

Original · left

Sampling (music)

Reusing a portion of an existing recording in a new song.

First in use: the 1970s

Duplicate · right

Sampling (language models)

Choosing each next token from a model's probability distribution.

First in use: c. 2018

Sampling / Sampling (music) / Sampling (language models)

Loading data…

Canonical todayMeasuring…

Status—

Duplicate share · 12 mo—

Last crossover—

Colour keyGrey · original meaningGreen · duplicate meaningAmber · contested (40–60%)How colours work

Hip-hop samples vs. top-p.

Which meaning of Sampling is the original, and which one is the duplicate riding on its name? Sampling (music) came first (the 1970s). Sampling (language models) arrived in 2018. The data below shows which one owns the word now.

Wikipedia pageviews since 2015

Sampling (music)Sampling (language models)

English Wikipedia pageviews since July 2015 load here.

Duplicate's share of attention

Green line = duplicate's % · above 60 it owns the word · below 40 the original does

The case for each meaning

Sourced definitions, usage stats and quotes

Original · the 1970s

Sampling (music)

Loading the Wikipedia definition…

12-mo pageviews—

Year over year—

Peak month—

Duplicate · c. 2018

Sampling (language models)

Loading the Wikipedia definition…

12-mo pageviews—

Year over year—

Peak month—

More duplicated terms

AI & machine learning

Prompt · Prompt (AI)Hallucination · Hallucination (AI)Token · Token (language model)Attention · Attention (machine learning)Weight · Weight (neural network)Bias · Bias (machine learning)Temperature · Temperature (LLM sampling)Embedding · Embedding (AI)Transformer · Transformer (deep learning)Model · Model (AI)Training · Training (AI)Agent · AI agent Grounding · Grounding (AI)Alignment · AI alignment Loss · Loss function Epoch · Epoch (machine learning)Pruning · Pruning (neural networks)Distillation · Knowledge distillation