Computer Science Stack Exchange is a question and answer site for students, researchers and practitioners of computer science. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

From Wikipedia

Information retrieval is the activity of obtaining information resources relevant to an information need from a collection of information resources. Searches can be based on metadata or on full-text indexing.

From Wikipedia

Information extraction (IE) is the task of automatically extracting structured information from unstructured and/or semi-structured machine-readable documents. In most of the cases this activity concerns processing human language texts by means of natural language processing (NLP). Recent activities in multimedia document processing like automatic annotation and content extraction out of images/audio/video could be seen as information extraction.

What are the relations and differences between information retrieval and information extraction?


share|cite|improve this question

Information retrieval is based on a query - you specify what information you need and it is returned in human understandable form.

Information extraction is about structuring unstructured information - given some sources all of the (relevant) information is structured in a form that will be easy for processing. This will not necessary be in human understandable form - it can be only for use of computer programs.

Some sources:

share|cite|improve this answer gives a very nice, concise distinction:

Information Extraction is not Information Retrieval: Information Extraction differs from traditional techniques in that it does not recover from a collection a subset of documents which are hopefully relevant to a query, based on key-word searching (perhaps augmented by a thesaurus). Instead, the goal is to extract from the documents (which may be in a variety of languages) salient facts about prespecified types of events, entities or relationships. These facts are then usually entered automatically into a database, which may then be used to analyse the data for trends, to give a natural language summary, or simply to serve for on-line access.

To put it pictorially:

Information Retrieval gets sets of relevant documents:

enter image description here

Information Extraction gets facts out of documents:

enter image description here

share|cite|improve this answer

From a modeling standpoint, information retrieval is a deep field predicated on several disciplines, including statistics, math, linguistics, artificial intelligence and now data science. In practice, these models are applied against text within corpora to discover patterns in the data. Not only do the IR models overlap in their usage, they can "partner" with other models such as k-means or k-nearest neighbor models, then other models can be applied from the vantage point of computational linguistics such as LDA/LDI and topic modeling Then, the end game is some sort of information visualization of this discovery--after ranking, clustering and aggregating work. Information Retrieval may appear to be a cryptic discipline, but serious effort, which is greatly appreciated, is going into opening up the area for deeper understanding of each model, and the interaction between models. I cite "Synthesis Lectures on Information Concepts, Retrieval, and Services" Series as the best place to delve into a foundation for IR.

While I don't entirely separate IR and Information Extraction, perhaps a subset of IE, concept level extraction, does apply IR patterns along with AI-based inferencing rules to extract related ontologies. The graphical nature of these relations are being enhanced with ontology modeling in OWL and RDF, and with graph databases, which allow for a less strict or rigorous set of relationship modeling, and allow for more relationships to surface, rather than being controlled per se. The ability to grow information extraction dynamically keeps its "discipline" strongly interesting to researchers.

Both IR and IE play out in our own significant "entities of the moment"--some have called "dynamic ontologies"--some being Palantir-- we need the patterns, models, simulations and visualizations of those significant entities to do business in the face of morphing new sources of information and changing of existing information. The conceptual, relational, definitional, pattern and ontological modeling have to be flexible and their visualizations the same. The heavy lifting of AI engines such as Watson in the information extraction and inferencing fields has cast a spotlight on the IE and frankly IR fields. Also the ubiquity of natural language processing and machine learning are calling attention to IR and IE models and engines. The impact of IR models on search and SEO, and on semantic web modeling is one of those "watching a change agent change due to impacts back from what it impacted" events--somewhere in harmonics and relativity theory. IR and IE are expanding based on what they're impacting.

share|cite|improve this answer

Information retrieval is about returning the information that is relevant for a specific query or field of interest. Note that this information could also be in the form of general documents, sure enough search engines are a notable example of such task. I would say that the most important entities recognizable for information retrieval are the initial set of documents/information and the query that specify "what to search for".

On the other hand information extraction is more about extracting (or inferring) general knowledge (or relations) from a set of documents or information. Note that here all the content of the documents could be considered as a whole corpus of data from which extract the knowledge. Of course also for this case you can somehow specify what do you want to extract, but it is more about properties/relations than specific subjects/topics. Properties are more domain-specific, while generally relations cover more generic scenarios.

Again, with search engines you're asking to get the sites that are most likely to contain information about that specific subject. This is an example of information retrieval.

For information extraction you could instead, for example, ask to extract all the names of cities, or e-mail addresses, that appear in a corpus of documents. You could even go much more generic, asking simply to extract knowledge. As you can see this is really generic, but it can be accomplish, for example, by obtaining triplets of the form subject-action-object for each valid sentence of a text (this is best suited for natural language texts).

If you're interested these (and other) topics are explain in details in the Natural Language Processing chapter of the book Arti ficial Intelligence: A Modern Approach.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.