Winter School on Fairness, Accountability, and Transparency in Artificial Intelligence Hosted by IIT Kharagpur
Sponsored by IBM Research India
Dates: 11 to 22 January 2021
Course coordinators:
- Academic: Manjira Sinha, IIT Kharagpur[email protected] and Abhijnan Chakraborty, MPI-SWS/IIT Delhi [email protected]
Description of school:
In present times, a variety of social sectors and services rely on algorithmic models to perform with optimal efficiency and to ensure maximum good. In a sense, artificial intelligence (AI) based systems have become more cross-disciplinary; a socio-technical and legal entity. However, algorithms are imperfect and they carry the similar biases and constraints prevalent in our society and culture. Hence it is important to investigate and comprehend the limitations of the machine learning models and make them more fair, transparent and accountable— in a word, more ethical. This school attempts to address some fundamental and pressing topics in the domains of fairness, accountability and transparency in AI.
We strongly encourage female candidates to apply. Special consideration will be given to female applicants.
List of subtopics:
- Introduction to ML
- Bias in online systems, pillars to trust/FAT
- Fairness in ML:
- Classification, regression, clustering
- Search ranking, recommendations
- Explainability:
- Overview, motivation and examples
- Detailed algorithms of local and global explainability
- Tutorial - AIX360 which will include Lime and Shap
- Explainability in real world
- AI for society:
- Abuse, hate speech and counter-speech detection
- Fake news detection, mis/disinformation online
- Tackling mental health issues
- Audits and lineage
Proposed list of speakers:
- Niloy Ganguly, CSE, IIT Kharagpur
- Animesh Mukherjee, CSE, IIT Kharagpur
- Pawan Goyel, IIT Kharagpur
- Saptarshi Ghosh, IIT Kharagpur
- Sourangshu Bhattacharya, IIT Kharagpur
- Abhijnan Chakraborty, MPI-SWS/IIT Delhi
- Pranay Lohia, IBM Research India
- Vijay Arya, IBM Research India
Background/prior courses recommended:
- Course will be open to undergraduate students from CS/IT stream; students must be in sixth semester or above, during the course.
- Basic knowledge about algorithms, machine learning and artificial intelligence is required.