cyberquantic logo header
EN-language img
FR-language img
breadcrumbs icon
Customer Relationship

Self-Service Customer Support

Better human-computer interaction with advanced language models

Better human-computer interaction with advanced language models

For:
End users
Goal:
Other
Problem addressed
Improve the quality of human-computer interaction.
Scope of use case
Human-computer interaction
Description
Natural language processing (NLP) technologies are crucial
for interaction of social systems and artificial intelligence
algorithms. AI models used in NLP are typically trained on
large amounts of text (corpora) in order to make them
learn language in general (language models). Then these
models are fine-tuned to particular down-stream tasks like
question answering, paraphrasing, fake news detection, etc.
High quality models require large amounts of data to train
and thus large computer systems are needed. Additionally,
we employ neural architecture search and automated
hyperparameter optimization techniques to derive better
models. Finally, we built a rigorous evaluation framework to
explore how model architectures and hyperparameters,
including source corpora and pre-processing methods, affect
performance of the models on each of the down-stream
tasks.
act frame img
Conversational agents
Interested in the same or similar project?
Submit a request and get a free project evaluation.