Biplav Srivastava  Biplav Srivastava photo       

contact information

Senior Researcher, IBM Master Inventor
IBM Research, New Delhi, India
  +91dash9810dash404142

links

Professional Associations

Professional Associations:  ACM  |  Association for the Advancement of Artificial Intelligence (AAAI)  |  IEEE


Many have approached me for research ideas. With that motivation, here are some ideas for research projects suggested with the aim to promote high quality research that is globally useful.

Gandhi ji's talisman

"I will give you a talisman. Whenever you are in doubt, or when the self becomes too much with you, apply the following test. Recall the face of the poorest and the weakest man [woman] whom you may have seen, and ask yourself, if the step you contemplate is going to be of any use to him [her]. Will he [she] gain anything by it? Will it restore him [her] to a control over his [her] own life and destiny? In other words, will it lead to swaraj [freedom] for the hungry and spiritually starving millions? Then you will find your doubts and your self melt away."


Problem #1

Statement:

Suggest public health plan for a seasonal disease in a future period based on open data

Description:

There are many seasonal diseases - dengue, malaria, Japanese echephalitis, ... There are many known techniques to tackle them - put health advertisement, de-fogging, vaccines, mosquito nets, ... which come at different costs and relative effectiveness. Open data from government, like data.gov.in's site and APIs, provide data for many years on them. Still, in 2014, starting with May and monsoon, hundreds of people will die. What public health plans can we put in as advisories for all districts of India on what to do and when, and how many disease cases/ diseases will it reduce? The default public health plan is to do the same as last years, i.e., nothing, and have linear increase in reported cases and deaths.

The technical challenges will be (1) data related: missing data, aggregate data, .. (2) selecting right analytical methods that account for data issues, and (3) generating cost-efficient public health plans.

Problem #2

Statement:

Create a travel recommendation system to suggest travel plans for individuals that reduces footprint on the road

Description:

Travel recommenders suggest ways for an individual to travel in a city including one or more information from - path, mode, vehicle and time. However, they should also suggest it in a manner that the travel have least impact on the city's resources. Example, take less road space, not cause delays, need little route management. This is because both private and public resources need to be optimised for a viable long-term solution.

See further details here: Biplav Srivastava, Raj Gupta, Nirmit Desai (2014). "Challenge: Loosely Synchronized Multi-Modal Plans for Traffic Improvement and Commuter Convenience" in IBM Research Report No. RI14001, At: http://domino.research.ibm.com/library/cyberdig.nsf/index.html

Problem #3

Statement:

Create visualisations showing return of investment in public service domains - e.g., health, energy, crime, water, employment generation, and their relative efficiency

Description:

Conventional wisdom says that the more money you put in a process, the better the result should be. For different public service domains, result translates to different metrics but the principles stay the same. So, show how much was the money (resource) spent in an area and whether the corresponding result was obtained. Further, which area gave the best result. Example of metrics for different domains are: health - number of deaths per 1000, energy - annual hours of power outage per capita, crime - number of deaths per 1000 population, water - number of people with potable water per capita, employment - number of unemployed per 1000.