# Dane dla AI

EN: **Data for AI** · id `data`

Path: [AI Atlas](https://aiatlasgraph.com/llms.txt) › Dane dla AI

## Description

- PL: Pozyskiwanie, przygotowanie, selekcja i kontrola danych używanych przez modele.
- EN: Sourcing, preparing, selecting and controlling the data models are built on.

## In one sentence

- PL: To przygotowanie informacji i przykładów, z których modele uczą się i na których są sprawdzane.
- EN: This prepares the information and examples that models learn from and are tested on.

## Problem it solves

- PL: Błędne, przypadkowe lub niepełne przykłady mogą nauczyć model złych nawyków.
- EN: Incorrect, irrelevant or incomplete examples can teach a model the wrong patterns.

## What it does

- PL: Zbiera materiały, sprawdza ich pochodzenie, poprawia błędy i wybiera przykłady odpowiednie do nauki oraz osobnego sprawdzianu.
- EN: It collects material, checks its origins, fixes problems and selects examples for learning and a separate assessment.

## Everyday example

- PL: Przygotowujesz zdjęcia produktów do rozpoznawania uszkodzeń: oznaczasz wady, usuwasz powtórzenia i odkładasz część zdjęć do sprawdzenia modelu.
- EN: You prepare product photos for damage detection, label defects, remove duplicates and set some photos aside to test the model.

## This concerns you if you…

- PL: chcesz nauczyć model pracy na własnych materiałach
- PL: zastanawiasz się, skąd biorą się błędy modelu
- EN: want a model to learn from your own material
- EN: wonder where a model's mistakes come from

## Keywords

web crawling · data provenance · data cleaning · quality filtering · data annotation · weak supervision · data augmentation · synthetic data generation · dataset versioning · data mixtures · train-validation-test splits · data leakage

## How it works

- PL: Proces łączy pozyskiwanie próbek, nadawanie etykiet i transformacje z kontrolą pochodzenia. Podziały zbioru i pomiary jakości pozwalają sprawdzić, czy rozkład danych odpowiada zadaniu i warunkom użycia modelu.
- EN: The process combines acquisition, labelling and transformations with provenance tracking. Dataset splits and quality measurements test whether the data distribution fits the task and the model's intended operating conditions.

## Limits

- PL: Większy zbiór nie kompensuje błędnych etykiet ani brakujących przypadków. Koszt obejmuje anotację, przechowywanie, prawa do użycia i utrzymanie jakości podczas zmian źródeł.
- EN: More data cannot compensate for incorrect labels or missing cases. Costs include annotation, storage, usage rights and maintaining quality as sources change.

## Key terms

- PL: **rozkład danych**: Częstości i zależności przypadków występujących w zbiorze.
- PL: **pochodzenie**: Informacja o źródłach próbek i wykonanych na nich operacjach.
- PL: **etykieta**: Odpowiedź referencyjna przypisana przykładowi podczas przygotowania danych.
- EN: **data distribution**: Frequencies and relationships of cases occurring in a dataset.
- EN: **provenance**: Information about sample origins and operations performed on them.
- EN: **label**: A reference answer assigned to an example during data preparation.

## Know first

- [Uczenie nadzorowane (Supervised learning)](https://aiatlasgraph.com/t/supervised.md)

## Examples

- PL: Crawling · filtrowanie jakości · mieszanki danych · etykietowanie · dane syntetyczne
- EN: Crawling · quality filtering · data mixtures · labelling · synthetic data

## Subtopics (9 direct, 48 in total)

- [Atrybucja i wycena danych treningowych (Training data attribution and valuation)](https://aiatlasgraph.com/t/datavaluation.md): Methods estimating the influence of training examples on predictions or their value for model quality, including influence functions, TRAK and Shapley values.
  - [Funkcje wpływu i śledzenie przykładów treningowych (Influence functions and training example tracing)](https://aiatlasgraph.com/t/influencefunctions.md): Gradient-based approximations of a training example's influence on a test prediction: classical influence functions, TRAK and scalable variants.
  - [Wycena danych wartością Shapleya (Data Shapley) (Data Shapley valuation)](https://aiatlasgraph.com/t/datashapley.md): The cooperative-game Shapley value applied to training examples, estimated by Monte Carlo sampling or gradient methods.
- [Czyszczenie i jakość (Data cleaning and quality)](https://aiatlasgraph.com/t/cleaning.md): Removing errors, duplicates and unwanted examples.
  - [Deduplikacja i dekontaminacja (Deduplication and decontamination)](https://aiatlasgraph.com/t/dedup.md): Removing duplicates and near-duplicates, and cutting training examples that overlap with test sets.
    - [Deduplikacja semantyczna (po embeddingach) (Semantic deduplication with embeddings)](https://aiatlasgraph.com/t/semanticdedup.md): Removing semantic near-duplicates by clustering embeddings from a pretrained model and pruning pairs above a similarity threshold.
  - [Ekstrakcja tekstu ze stron i PDF (Text extraction from HTML and PDFs)](https://aiatlasgraph.com/t/textextraction.md): Pulling the actual content out of raw pages and documents: removing menus, footers and ads, turning PDFs (including scans) into clean text while keeping structure.
  - [Filtrowanie jakości (heurystyki i klasyfikatory) (Quality filtering with heuristics and classifiers)](https://aiatlasgraph.com/t/qualityfiltering.md): Dropping weak documents before training — first with simple rules, then with a model that scores how valuable a text is (e.g. educationally).
  - [Filtrowanie szkodliwych treści tekstowych w danych treningowych (Harmful text filtering in training data)](https://aiatlasgraph.com/t/contentsafetyfiltering.md): Classifier- and rule-based removal of toxic, hateful and sexual content from text pretraining corpora.
  - [Niezbalansowane klasy i wartości odstające (Class imbalance and outliers)](https://aiatlasgraph.com/t/classimbalance.md): Spotting and handling data where one class dominates or unusual examples appear, before they distort training and evaluation.
  - [Wykrywanie błędnych etykiet (Label error detection)](https://aiatlasgraph.com/t/labelerrors.md): Automatically finding examples that probably carry a wrong label, using model predictions, instead of reviewing the whole dataset by hand.
- [Dane syntetyczne i augmentacja (Synthetic data and augmentation)](https://aiatlasgraph.com/t/synthetic.md): Creating or transforming examples.
  - [Augmentacja danych (obraz i tekst) (Data augmentation for images and text)](https://aiatlasgraph.com/t/classicaugmentation.md): Creating new examples by transforming existing ones — rotating and cropping images, paraphrasing and back-translating text.
  - [Generowanie danych instrukcyjnych przez LLM (LLM-generated instruction data)](https://aiatlasgraph.com/t/instructiongen.md): A model generates instructions, inputs and answers from a few seed examples; after filtering weak and near-duplicate ones they are used for fine-tuning.
  - [Syntetyczne dane do pretrainingu (przepisywanie, podręczniki) (Synthetic pretraining data: rephrasing and textbooks)](https://aiatlasgraph.com/t/pretrainsynthetic.md): Rewriting web text into a more useful form or generating textbook-style material as a way to improve pretraining efficiency. Outcomes depend on generation quality, filtering, and how synthetic data is combined with real data.
  - [Syntetyczne dane tabelaryczne (Synthetic tabular data generation)](https://aiatlasgraph.com/t/tabularsynthetic.md): Generating rows for mixed-type tables with generative models; CTGAN conditions generation to better represent multimodal continuous columns and rare categories.
  - [Syntetyczne trajektorie i dane wywołań narzędzi dla agentów (Synthetic agent trajectories and tool-use data)](https://aiatlasgraph.com/t/agenttrajectories.md): Automatic generation and verification of queries, function calls and multi-step agent trajectories in environments and APIs for model fine-tuning.
  - [Weryfikacja i filtrowanie danych syntetycznych (Verifying and filtering synthetic data)](https://aiatlasgraph.com/t/syntheticfiltering.md): Quality control for model-generated data through rejection sampling and automatic verification of answers, code execution or tool calls.
  - [Wzbogacanie próbek (Sample enhancement)](https://aiatlasgraph.com/t/dataenhance.md): Improving generated examples: checking answers, picking the best of several, making tasks harder.
  - [Załamanie modelu na danych syntetycznych (Model collapse from synthetic data)](https://aiatlasgraph.com/t/modelcollapse.md): The risk that models trained recursively on content generated by earlier models lose information about the original data distribution, beginning with rare cases.
- [Etykietowanie i feedback (Annotation and feedback)](https://aiatlasgraph.com/t/annotation.md): Producing reference answers and judgements.
  - [Aktywne uczenie (active learning) (Active learning)](https://aiatlasgraph.com/t/activelearning.md): The model picks which examples are worth labelling next, to reach good quality with fewer labels.
  - [Etykietowanie z pomocą LLM (LLM-assisted annotation)](https://aiatlasgraph.com/t/llmannotation.md): Using language models to produce labels, rationales and pairwise judgements, with quality checked on a human-reviewed sample.
  - [Słaba superwizja i funkcje etykietujące (Weak supervision and labeling functions)](https://aiatlasgraph.com/t/weaksupervision.md): Instead of labelling by hand, you write many simple, imperfect rules and a model combines their votes into a single better label.
  - [Zbieranie danych preferencji (porównania odpowiedzi) (Preference data collection (response comparisons))](https://aiatlasgraph.com/t/preferencedatacollection.md): Designing and running collection of pairwise comparisons or response rankings from people as data for reward models and preference optimisation.
  - [Zgodność anotatorów (Inter-annotator agreement)](https://aiatlasgraph.com/t/annotationguidelines.md): Measuring how consistently independent annotators assign labels while accounting for chance agreement. A low score requires diagnosis: it may reflect unclear categories, task subjectivity, class distribution, or an unsuitable agreement coefficient.
- [Kuracja danych multimodalnych (Multimodal data curation)](https://aiatlasgraph.com/t/multimodaldata.md): Filtering, segmenting and (re)captioning image–text, video and speech data before pretraining multimodal and generative models.
  - [Filtrowanie par obraz–tekst (Image–text pair filtering)](https://aiatlasgraph.com/t/imagetextfiltering.md): Selecting image–text pairs from a web pool by embedding similarity, rules and classifiers, judged by the quality of the trained model.
  - [Kuracja danych wideo (Video data curation for generative models)](https://aiatlasgraph.com/t/videocuration.md): A multi-stage pipeline: cut detection, clip captioning and filtering by motion, on-screen text, aesthetics and image–text alignment.
  - [Przygotowanie danych mowy z nagrań „z życia” (In-the-wild speech data curation)](https://aiatlasgraph.com/t/speechdatacuration.md): A pipeline of normalisation, source separation, diarization, VAD segmentation, automatic transcription and quality filtering for large speech corpora.
  - [Syntetyczne podpisy do obrazów (recaptioning) (Synthetic image recaptioning)](https://aiatlasgraph.com/t/recaptioning.md): Generating new, dense image descriptions with a vision-language model and training on them instead of or alongside original captions.
- [Pipeline’y i wersjonowanie (Data pipelines and versioning)](https://aiatlasgraph.com/t/pipelines.md): Repeatable data preparation across experiments.
  - [Przetwarzanie danych na dużą skalę (Large-scale data processing)](https://aiatlasgraph.com/t/largescaleprocessing.md): Building repeatable pipelines that read, filter, deduplicate and tokenise trillions of tokens in parallel across many machines.
  - [Strumieniowe, shardowane zbiory treningowe (Streaming sharded training datasets)](https://aiatlasgraph.com/t/streamingdatasets.md): Organising samples into sequentially read shards and streaming them from disk or object storage into training.
  - [Walidacja i testy danych (Data validation and testing)](https://aiatlasgraph.com/t/datavalidation.md): Automatically checking that incoming data has the expected schema, ranges and distributions before it reaches training or a production model.
  - [Wersjonowanie zbiorów danych (Dataset versioning)](https://aiatlasgraph.com/t/dataversioning.md): Keeping versions of data the way you keep versions of code, so every model can be tied to the exact dataset it was built on.
- [Podziały i wyciek danych (Data splits and leakage)](https://aiatlasgraph.com/t/splits.md): Keeping study material apart from an honest exam.
  - [Rodzaje wycieku danych i ich wykrywanie (Types of data leakage and their detection)](https://aiatlasgraph.com/t/leakagetypes.md): Recognising common mistakes that let a model peek at the answers: preprocessing train and test together, duplicates, features unavailable in practice, leakage from the future.
  - [Strategie podziału: losowy, czasowy, grupowy (Split strategies: random, temporal, grouped)](https://aiatlasgraph.com/t/splitstrategies.md): Choosing how to split data for the situation: by time for forecasting, by user or patient when examples are related to each other.
- [Pozyskiwanie i pochodzenie (Data sourcing and provenance)](https://aiatlasgraph.com/t/sourcing.md): Where the data comes from and whether we may use it.
  - [Licencje i zgody na użycie danych (Data licensing and consent)](https://aiatlasgraph.com/t/datalicensing.md): Checking the terms under which data may be collected and used for model training: dataset licences, terms of service, consent, and opt-out signals. Robots.txt rules govern crawler access but are not themselves a licence or legal authorization to use the data.
  - [Pozyskiwanie danych z sieci (crawling) (Web crawling for training data)](https://aiatlasgraph.com/t/webcrawl.md): Collecting web pages at scale as raw training material: using existing web snapshots or running your own crawler that respects site rules.
  - [Wyszukiwanie i ocena gotowych zbiorów danych (Finding and assessing public datasets)](https://aiatlasgraph.com/t/datasetdiscovery.md): Using dataset hubs, catalogues and their metadata (dataset cards, Croissant) to select, inspect and load data.
  - [Śledzenie pochodzenia danych (lineage) (Data lineage and traceability)](https://aiatlasgraph.com/t/datalineage.md): Recording where each part of a dataset came from, who created it and how it was processed, so it can be audited, fixed or removed.
- [Selekcja i proporcje danych (Data selection and mixing)](https://aiatlasgraph.com/t/dataselection.md): Deciding which examples, and in what proportions, enter training to improve model quality or budget efficiency when data sources differ in quality and usefulness.
  - [Dobór mieszanki danych (DoReMi, RegMix) (Data mixture optimisation)](https://aiatlasgraph.com/t/datamixing.md): Setting how much each source (web, code, books, science) contributes to training, e.g. using small proxy models.
  - [Kolejność danych i uczenie z programem nauczania (curriculum learning) (Data ordering and curriculum learning)](https://aiatlasgraph.com/t/datacurriculum.md): Controlling the order and pace of data: curricula from easier to harder examples and automatic self-paced learning.
  - [Selekcja danych do dostrajania (Selecting data for fine-tuning)](https://aiatlasgraph.com/t/finetuneselection.md): Choosing a small, carefully curated set of SFT examples instead of a large random one — by hand or with a method that measures each example's influence on the target task.

## Sources

- [Full Stack Deep Learning (2022, archiwum)](<https://fullstackdeeplearning.com/course/2022/>) `fsdl` · reference: Pełny cykl systemu ML: dane, testy, wdrażanie, monitoring, zespoły i etyka.
- [Stanford CS336 · Language Modeling from Scratch (2025)](<https://cs336.stanford.edu/spring2025/>) `cs336` · reference: Budowa modeli, dane, trening, systemy, inferencja i post-training.
- [Introduction to Data-Centric AI (MIT IAP course, 2023–2024)](<https://dcai.csail.mit.edu/>) `dcai-mit` · by MIT CSAIL (prowadzący na stronie tylko z imion: Anish, Curtis, Jonas)
- [A Survey of LLM × DATA](<https://arxiv.org/abs/2505.18458>) `llm-x-data-survey-2025` · by Xuanhe Zhou, Junxuan He, Wei Zhou, Haodong Chen, Zirui Tang, Haoyu Zhao, Xin Tong, Guoliang Li i in. (17 autorów)

---

AI Atlas revision 122 (2026-10-30) · HTML: https://aiatlasgraph.com/t/data · index: https://aiatlasgraph.com/llms.txt · everything: https://aiatlasgraph.com/llms-full.txt · updated 2026-10-30