Interactive Machine Learning
Interactive Machine Learning (IML) is the design and implementation of algorithms and intelligent user interface (IUI) frameworks that facilitate machine learning (ML) with the help of human interaction.
With the convergence of Artificial intelligence and Machine Learning, IML is where the Human-Computer Interaction (HCI) and IUI community meets the ML community. We also combine contributions from related fields such as data science, cognitive science, computer graphics, design or the arts, and natural language processing, data mining/data analytics, and knowledge representation and reasoning. Our focus is to improve the interaction between humans and machines to update ML models, by leveraging both state-of-the-art HCI and IUI approaches, as well as solutions that involve state-of-the-art ML techniques.
We can “assist” AI systems in becoming self-sustaining, “lifelong” learners. This includes to (1) develop insights into the importance of the social and cultural contexts of machine learning; (2) create ML systems that actively seek information; (3) realise the need to pay attention to the incomplete context understandings and naive generalisations that ML systems, in particular end-to-end systems, bring with them to a given subject. We focus on multimodal multisensor interfaces for ML.
Topics
Mixed initiative interaction:
The ML system and the domain expert engage in a two-way dialogue to facilitate more accurate learning from less data compared to the classical approach of passively observing labeled data.
- Crowdsourcing
- Incremental learning
- Integral / One-shot learning
- Multimodal Multisensor Interfaces
- Online learning
- Speech-based ML
- Textual entailment for ML
- Argumentative active learning (from textual explanation feedback)
Transparency and human interpretability in ML:
- Mitigate bias in ML
- Evaluate machine learning models
- Render ML models more interpretable
- Interpretation of black-box models (including deep neural networks)
- Enhance trust in ML
- Causality of predictive models
- Visual analytics
Explainable artificial intelligence applications:
- Medical decision making
- Reliance on learned models in deployed applications
- Legal requirements
General AI for ML:
- Metalevel verification and validation
- Metacognition and introspection
- Super-human AI Reasoning
- Reinforcement learning
- Active Learning
- Subgroups
IML tools for enhancing human cognitive capabilities:
- Help understand and foresee long-term implications of ML
- Empower humans to make important decisions in a more informed way
To fully automate tasks can be extremely difficult or even undesirable: With IML, we can increase the reach of ML solutions for end users by using IUIs for interactively learning and applying ML models. This includes visuals analytics and visual data mining, elegant knowledge elicitation (active and passive) by multimodal-multisensor interfaces, collaborative learning, and adaptive and adaptable data analysis according to user models. User input is used to choose, develop or criticise models, algorithms, their hyper-parameters, or the multimedia training data themselves. User input provides (task-specific) domain knowledge and constraints.
Project Categories