VolgendeVolgende gids
LaMa Resolutie-Robuust Inschilderen
Visuele AI
Taal AI-GIDS
Coreference resolution links mentions in a text that refer to the same entity, such as a name and a pronoun.
Luister naar deze gidsAI-verteld · ~3 min
This supports tracking who or what is being discussed, while ambiguous references may require context or remain unresolved.
Texts often introduce an entity by name and refer to it later through pronouns or descriptions. Coreference resolution groups mentions that identify the same entity in the discourse. In Maria said she was tired, a system can investigate the link between Maria and she; a different context may require a different interpretation. Resolving a pronoun back to an earlier referring expression is a familiar form of pronominal anaphora. Not every relation between expressions is identity coreference, so the task should be defined carefully. Winograd-schema examples illustrate a difficulty beyond grammatical agreement. In the trophy and suitcase example, knowing what cannot fit when it is too large helps interpret the pronoun. Grammar alone permits both noun phrases as candidates. Incorrect mention links can also affect a summary or answer: confusing a company with its product can change who performed an action. Influential span-ranking models score candidate mentions and potential earlier antecedents, constructing clusters through predicted links. This is a description of a model family, not every possible coreference algorithm. Distance, agreement, and contextual information can contribute evidence, without making a surface cue definitive. Entity linking adds another task: associating a textual mention with an entry in a knowledge base. Grouping Maria and she within a document does not necessarily identify the relevant external database record. Evaluation should also check biases. WinoBias contains occupation-based, Winograd-style examples that compare pro-stereotypical and anti-stereotypical pronoun associations. It examines whether stereotype-consistent choices receive an advantage in the evaluated systems rather than assuming occupation determines gender.
Checking whether a summarizer keeps a person's name and later pronouns associated correctly.
Comparing a textual mention cluster with a link to an external knowledge-base entry.
Investigating a QA error caused by linking a company pronoun to a product.
Evaluating occupation-based pronoun examples for stereotypical bias.
Reference tracking can be evaluated in longer documents, dialogue, and multilingual text, with attention to the annotation conventions used. Linking mentions across documents introduces additional identity ambiguities, while knowledge-base linking requires external candidate information. Comparisons should distinguish those tasks rather than treat every entity-matching score as a coreference result. A benchmark can also reveal unequal treatment of different pronoun or occupation examples. Clear evidence, appropriate uncertainty, and bias evaluation help interpret a model's links without treating correct resolution of a few pronouns as proof of complete language understanding.
Gratis nieuwsbrief
Drie verifieerde AI-verhalen elke weekdagmorgen, geschreven in eenvoudig Engels. Gratis voor altijd, geen advertenties.
Eén e-mail elke weekdag. Afmelden met één klik. We verkopen of delen je adres nooit.
Test jezelf
Directe feedback op elke antwoord, en een deelbare certificaat met een verifieerbare ID zodra je een cursus hebt afgerond.
Ondersteun gratis AI-onderwijs. AI Understanding is een 501(c)(3) non-profit — geen advertenties, geen betaalmuur, nooit. Doe een donatie
Coreference resolution links mentions in a text that refer to the same entity, such as a name and a pronoun. This supports tracking who or what is being discussed, while ambiguous references may require context or remain unresolved.
Identity coreference groups mentions that refer to the same entity in discourse.
Resolving a backward-referring pronoun is a form of pronominal anaphora; not every anaphoric relation is identity coreference.
Zowel 'trofee' als 'koffer' zijn grammaticaal geldige antecedenten voor 'het'; Om correct te kiezen, moet je weten dat iets dat te groot is, niet zal passen, namelijk wereldkennis.
Als een systeem niet kan vaststellen dat 'het' het bedrijf betekent en niet het product, kunnen de antwoorden en samenvattingen verderop in het proces feitelijk onjuist zijn.
This describes a span-ranking family rather than a universal algorithm for all coreference systems.
Blijf leren
Er zijn meer handleidingen voor dit onderwerp geselecteerd
VolgendeVolgende gids
LaMa Resolutie-Robuust Inschilderen
Visuele AI