IBM and IIT-Delhi partner for AI research

IBM and IIT Delhi have partnered to carry out research in artificial intelligence. IIT Delhi will join IBM’s AI Horizons Network for a multi-year research collaboration to help organizations take decisions with the help of AI systems, reports the Economic Times. They will use natural language techniques to gain (unspecified) domain knowledge.

Michael Karasick, VP of Global Labs at IBM Research, said that the collaboration will help organizations make informed decisions by ‘infusing reasoning, comprehension and transparency in their AI systems.’

In September, IIT-Bombay and IBM partnered for AI research as well, and IIT-Bombay joined the AI Horizons Network. The department of computer science and engineering at IIT-Bombay teamed with IBM Research to advance the applications of AI, machine learning, natural language processing in business and industries, such as in healthcare, financial services and retail, among others.

IIT-Kharagpur studies Twitter posts from post-disaster situations

  1. Microblog Retrieval for Post-Disaster Relief: Applying and Comparing Neural IR Models
    • This study was conducted by IIT-K with support from Microsoft Research India, Information Technology Research Academy (ITRA), DeITY (Government of India). It attempted to identify tweets in post-disaster situations in order to aid disaster relief and resource needs.
    • The study compared different methods for retrieving two kinds of tweets in post-disaster responses – need-tweets and availability-tweets. It used datasets of tweets posted during two disaster events – the 2015 Nepal earthquake and the 2016 Italy earthquake.
    • The study compared pattern matching techniques, language model based techniques and several neural IR (information retrieval) models to test which method was most successful in identifying the relevant tweets. The study found that neural IR models perform better than pattern-matching techniques, and also explored the possibility of reusing a pre-trained neural IR model to minimize training time with new datasets.
  2. Analysis and Early Detection of Rumors in a Post Disaster Scenario
    • Another study supported by ITRA and DeitY focused on identification of rumor tweets at an early stage after a disaster had occured. A model was developed by combining features of tweets which propagated rumours.
    • The model was tested over a large set of tweets after the 2015 Chennai Floods. The study observed that around 70% of the rumors were detected by the model, better than the other four tested models. It also achieved 0.99 point accuracy for the disaster scenario.

Two other studies were also carried out in the area of fake news and rumours:

  • Characterizing and Countering Communal Microblogs During Disaster Events
  • Extracting and Summarizing Situational Information from the Twitter Social Media during Disasters

India and Russia to invest $30 billion in AI and blockchain by 2025

India and Russia will increase cooperation in using blockchain and artificial intelligence, reports Inc42. The two countries released a statement after the first India-Russia Strategic Economic Dialogue was held last week at St Petersburg. The two countries have agreed to increase the two-way investment to $30 billion by 2025.

The dialogue focussed on five areas – transport infrastructure, agriculture and agro-processing sector, small & medium business support, digital transformation and frontier technologies, and industrial and trade cooperation. The next round of economic dialogue will be held in July/August 2019.