Versions Compared

Old Version 5

changes.mady.by.user Raul Hanson

Saved on

compared with

New Version 6

changes.mady.by.user Raul Hanson

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Intelligentse otsingu leiad aadressilt: https://otsing.taltech.ee/

You can find the intelligent search on this URL: https://search.taltech.ee/

Intelligentne otsingu allikad

Itelligent search sources

Kuidas lahendus töötab?

  1. Kasutaja avab rakenduse veebilehitsejas,

  2. Kasutaja peab sisse logima oma UNI-ID kontoga.

  3. Kasutaja esitab oma küsimuse otsingu lahtrisse

  4. Kasutaja küsimusest lähtuvalt rakendus otsib kõige relevantsemaid tekstilõike ülikooli algallikatest. Selleks kasutatakse OpenAI text-embedding-ada-002 mudelit.

  5. Tekstilõigus koos küsimustega edastakse GPT keelemudelile, mis genereerib lõpliku vastuse. Vaata mudelite kirjeldusi järgmises sektsioonis.

  6. Lõplik vastus koos algallikatega tagastakse rakenduses kasutajale.

  7. Jäta tagasisde tehisaru poolt antud vastuse kohta, et saaks paranda ja täiendada infot algallikates kui ka teha lahendust paremaks.

How does the solution work?

  1. The user opens the application in a web browser.

  2. User must log in with a UNI-ID account.

  3. The user submits his question in the search box

  4. Based on the user's question, the application searches for the most relevant text passages from the university's original sources. The OpenAI text-embedding-ada-002 model is used for this.

  5. The text passage with the questions is passed to the GPT language model, which generates the final answer. See model descriptions in the next section.

  6. The final answer with the original sources is returned to the user in the application.

  7. Leave feedback on the answer given by the artificial branch, so that the information in the original sources can be corrected and supplemented, as well as to make the solution better.

Mudelid

  • OpenAI GPT-3.5 Turbo on nagu nutikas tööriist, mis aitab sul leida erinevaid vastuseid oma küsimustele.

    • Mudel on võrdlemisi kiire, kuid annab natukene lihtsaid või mitte liiga sisukaid vastuseid.

  • OpenAI GPT-3.5 Turbo 16k on sarnane eelmisele, kuid suudab mõista rohkem infot korraga.

    • Mudel on võrdlemisi kiire, kuid annab natukene lihtsaid või mitte liiga sisukaid vastuseid.

  • OpenAI GPT-4 on veelgi nutikam ja teadlikum. See suudab mõista keerulisemaid probleeme ja teemasid. See on tööriist, mis on väga teadlik paljudest erinevatest teemadest.

    • Mudel annab häid ja sisukaid vastuseid, kuid vastuste genereerimine mudeli poolt hetkel võtab aega ning kasutajana tuleb olla kannatlik.

  • OpenAI GPT-4 Turbo on kõige uuem ja võimekam mudel. See suudab mõista ja meeles pidada veelgi rohkem infot. See on tööriist, mis on väga teadlik ja on samal ajal ka säästlik ning kuluefektiinve.

    • The model gives good and meaningful answers, but the generation of answers by the model takes time at the moment, and as a user you have to be patient.

  • Text-embedding-ada-002 - osingu (sh rühmitamise, soovituste, anomaaliate tuvastamise ja klassifitseerimise ülesande) jaoks kasutatakse “embeddings” mudelit, mis viib ülikooli põhiste allikate teksti numbrilisele kujule ning mis võimaldab mõõta ja siduda tektiosade vahelisi seoseid.

Täpsemalt mudelite kohta saab vaadata siit: https://platform.openai.com/docs/models/overview

Models

  • OpenAI GPT-3.5 Turbo is like a smart tool that helps you find different answers to your questions.

    The model is relatively fast, but gives somewhat simple or not very meaningful answers

    .

  • OpenAI GPT-3.5 Turbo 16k is similar to the previous one, but can understand more information at once.

    • The model is relatively fast, but gives somewhat simple or not very meaningful answers.

  • OpenAI GPT-4 is even smarter and more aware. It can understand more complex problems and topics. It is a tool that is very knowledgeable about many different topics.

    The model gives good and meaningful answers, but the generation of answers by the model takes time at the moment, and as a user you have to be patient

    .

  • OpenAI GPT-4 Turbo is the newest and most capable model. It can understand and remember even more information. It is a tool that is very knowledgeable and at the same time economical and cost effective.

    • The model gives good and meaningful answers, but the generation of answers by the model takes time at the moment, and as a user you have to be patient.

  • Text-embedding-ada-002 - embeddings model is used for the search (including clustering, recommendation, anomaly detection and classification tasks), which brings the text of university-based sources into a numerical form and which allows to measure and relatedness the relationships between text parts.

More details about the models can be found here: https://platform.openai.com/docs/models/overview

Seaded

Mudeli valimine

  • OpenAI GPT-3.5 Turbo - mudel on võrdlemisi kiire, kuid annab natukene lihtsaid või mitte liiga sisukaid vastuseid.

  • OpenAI GPT-3.5 Turbo 16k - mudel on võrdlemisi kiire, kuid annab natukene lihtsaid või mitte liiga sisukaid vastuseid.

  • OpenAI GPT-4 - mudel annab häid ja sisukaid vastuseid, kuid vastuste genereerimine mudeli poolt hetkel võtab aega ning kasutajana tuleb olla kannatlik.

  • OpenAI GPT-4 Turbo - mudel annab häid ja sisukaid vastuseid, kuid vastuste genereerimine mudeli poolt hetkel võtab aega ning kasutajana tuleb olla kannatlik.

Algallikad

Settings

Choosing a model

OpenAI GPT-3.5 Turbo - the model is relatively fast, but gives somewhat simple or not very meaningful answers.

OpenAI GPT-3.5 Turbo 16k - the model is relatively fast, but gives somewhat simple or not very meaningful answers.

OpenAI GPT-4 - the model gives good and meaningful answers, but the generation of answers by the model takes time at the moment, and the user has to be patient.

OpenAI GPT-4 Turbo - the model gives good and meaningful answers, but the generation of answers by the model takes time at the moment, and the user has to be patient.

Sources

Piiranugd ja edasised arendused

Funktsionaalsus

  • Puudub vestluse (chat) funktsionaalsus, tehisaru ei jäta meelde su eelnevat küsimsut ja ei oska selle pealt edasisi täpsutusi teha. Tulevikus soovime selle lisada.

  • Tulevikus soovime uuendada “embeddings” mudelit text-embedding-ada-002 uuemate ja võimekate vastu. See võimaldaks leida algselt küsija küsimuse peale meie algallikate seast veelgi täpsemad ja paremad vastused.

Plaanis tulevikus lisada juurde algallikatena:

Limitations and backlog of developments

Functionality

  • There is no chat functionality, the artificial intelligence does not remember your previous question and cannot make further clarifications based on it. We would like to add it in the future.

  • In the future, we would like to upgrade the current embeddings model text-embedding-ada-002 to newer and more capable one. This would make it possible to find even more accurate and better answers to the question of the original asker from among our original sources.

We plan to add more as original sources in the future: