What-if analysis on intent detection models and datasets

Research a method to visualize the impact of changes in the dataset for an intent detection solution impacts the performance of the trained intent detection model, without training the model.

Required interest(s)

  • Responsible, explainable, and socially aware data science
  • Human-centered and community-minded information systems
  • Language processing

What do you get

  • A challenging assignment within a practical environment
  • € 1000 compensation, € 500 + lease car or € 600 + living space
  • Professional guidance
  • Courses aimed at your graduation period
  • Support from our academic Research center at your disposal
  • Two vacation days per month

What you will do

  • 65% Research

Intent detection models are used frequently to build chatbots, text analysis solutions, and knowledge mining solutions. There are great tools available to build a dataset for an intent detection model. However, these models don’t offer good tools to perform a what-if analysis on your intent detection dataset.

Users are often resorting to trial and error to improve the dataset for their intent detection. We want to research better methods of determining what data could have a positive effect on the performance of the intent detection model.

Ideally, we want to provide the user with an impact score for samples to let them understand the effect of adding a specific sample to the training dataset of the intent detection model.

Another interesting method of performing what-if analysis could be the application of model explainers so the user gets a clear understanding of what exactly is picked up from a sample that is added to the training dataset.

About Info Support Research Center

We anticipate on upcoming and future challenges and ensures our engineers develop cutting-edge solutions based on the latest scientific insights. Our research community proactively tackles emerging technologies. We do this in cooperation with renowned scientists, making sure that research teams are positioned and embedded throughout our organisation and our community, so that their insights are directly applied to our business. We truly believe in sharing knowledge, so we want to do this without any restrictions.

Sign up for this assignment

  • Geaccepteerde bestandstypen: docx, doc, txt, pdf.

Application procedure

1

Introductory meeting

Discuss (study) career, interests and ambitions and introduction Info Support.

2

Review

Assessment of professional knowledge and personality (capacity, competences and motives).

3

Selection interview

Deepen professional knowledge and personality.

4

The signing of a contract

Contract offer and invitation for drawing moments.

Other Mastertheses

graduation assignment

Detecting forged documents with neural networks

Detect forged documents using a neural network. That is what this assignment is all about. One of our customers, a large credit provider, now checks all applications manually and is looking for a solu…

graduation assignment

What-if analysis on intent detection models and datasets

Research a method to visualize the impact of changes in the dataset for an intent detection solution impacts the performance of the trained intent detection model, without training the model.

graduation assignment

Protecting computer vision models against tampering when using transfer-learning

Research a method to protect state-of-the-art computer vision models against tampering in transfer-learning scenarios.

graduation assignment

Automatic summary of conference talks based on video and audio content

Research a method to create highlight videos of conference talks based on video and audio content by using a deep learning model.