Recently Mr. Dominic Cummings, Chief Political Strategist for Mr. Boris Johnson, Prime Minister of the United Kingdom, wrote a blog suggesting the need to revamp the decision-making strategies employed in the UK. Primarily he focused on the need to explore and exploit state of the art tools and techniques from mathematics and physics. Mr. Cummings wants to create a set of experts consisting of physicists, mathematicians and data scientists to deal with the major social and economic issues of the UK. Specifically, he wants to hire new talent in order to revamp and rejuvenate the structure of the decision making. In his own words, “We want to hire an unusual set of people with different skills and backgrounds to work in Downing Street with the best officials, some as spads and perhaps some as officials. If you are already an official and you read this blog and think you fit one of these categories, get in touch”.  Mr. Cummings, as a part of the top administration of the United Kingdom, wants to implement transdisciplinary concepts and techniques to improve the current strategies of policymaking for the UK. 

Although this may seem wishful, Mr. Cummings substantiates his stand for applying these novel technologies for a nation’s strategy building exercise in his blog. He emphasizes the need to hire young talent in order to build a team helping the Prime Minister in strategic planning. Mr. Cummings cites a list of important publications in order to support his claim of exploring the novel and non-conventional tools. He suggests that the aspirants desirous of participating in this exciting task with No: 10, Downing Street should go through the publications that he has listed. Interestingly, the very first paper that Mr. Cummings has cited comes from India, a collaborative effort by Dr. E. A. Gopalakrishnan (Assistant Professor, Amrita Center for Computational Engineering and Networking (CEN), Amrita Vishwa Vidyapeetham) who did his Ph.D. with Prof. Sujith, Prof. R. I. Sujith (Department of Aerospace Engineering, IIT Madras,, Mr. Tony John (Ph.D. student, Georgia Institute of Technology) who did his master’s thesis with Prof. Sujith, Prof. Partha Sharathi Dutta (Associate Professor, Department of Mathematics, IIT Ropar) and Ms. Yogita Sharma (Post-Doctoral Researcher, UC Berkeley) who did her Ph.D. with Prof. Dutta. These Indian scientists applied the tools from nonlinear dynamics to develop robust precursory measures for detrimental transitions observed in canonical thermoacoustic systems. Although there existed early warning measures for predicting transitions in mathematical models, Dr. Gopalakrishnan and his collaborators were the first to prove the efficacy of the same in predicting transitions in real physical systems.

Dr. Gopalakrishnan commented “First of all we are really happy to see that our work is cited by the scientific advisor of UK in relation to the new policies of attracting the human resources to work on the pressing problems of United Kingdom. In fact, we are pleasantly surprised to note the reach and implication of our results. It gives us immense pleasure to observe that our results can pave the way for the young talent to solve issues of the highest importance”.

The findings by Dr. Gopalakrishnan and his collaborators in the physical system backed with the experimental evidence can be applied to other areas such as finance, medicine, text analytics, cybersecurity, etc., to develop precursory measures for detrimental transitions.

It is really interesting to note the fact these researchers have applied techniques such as complex networks to predict transitions in combustion systems and are successful in their attempt. They are in the process of applying machine learning and deep learning techniques to develop robust early warning measures for complex systems.

A Brief Synopsis of the Ideas in the Paper 

Many of the natural, social and physical systems exhibit interesting and intriguing transitions from one stable state to an alternate stable state termed as tipping, for a minor variation in any of the system parameters. Often such transitions result in undesirable consequences such as the transition of the fertile ecosystem to a barren one, the sudden collapse of the economy of a country or sudden emergence of large amplitude oscillations in gas turbine engines. The catastrophic nature of these transitions demands the need to develop robust early warning measures to predict these transitions. 

Early warning measures can be developed by exploiting the dynamics associated with these transitions. A system tends to respond slowly to the change of parameters as it approaches a transition, termed as Critical Slowing Down in the dynamical system theory. The effect of critical slowing down can be quantified by calculating the autocorrelation and variance of any of the observables of the system. 

In this work, robust early warning measures were developed for the catastrophic transitions observed in a Rijke tube which is a prototypical thermoacoustic system. We recorded the acoustic pressure signals from the system by carefully conducting laboratory experiments and we were able to predict the impending transition by calculating autocorrelation, variance and conditional heteroskedasticity of the recorded pressure time series.

Relevance of the Research in the Indian Context 

The predictive techniques described in our research can be utilized in areas such as finance, healthcare, power generation, ecosystem management, etc. Effective utilization of predictive techniques in healthcare is highly pertinent to India where we spent a considerable amount of money in healthcare. Similarly, it will be beneficial for investors in a growing economy if we can predict highly catastrophic transitions that can happen in financial markets using the available data. 

The generation of clean and green energy is one of the most important goals for a country like ours which needs eco-sensitive developmental activities. Most modern-day power generation systems heavily utilize renewable energy to achieve the goal of producing clean energy. The increased use of renewable energy brings forth problems related to stability which eventually may lead to blackouts. Early warning measures can be employed to predict these undesirable transitions in power grids. These universal measures capable of predicting impending transitions can also be effectively utilized to determine and control the detrimental changes that happen in our ecosystems. 

The techniques described in the paper are not just limited to the analysis of natural or physical systems. They can also be used to understand the changes that happen in social systems. Currently, India is going through a stage wherein lot of novel policies are implemented in various domains. It is highly pertinent to understand the impact of these strategic decisions before and after they are implemented. In order to understand the impact of a state policy before it is completely implemented, various tools of predictive analytics can be employed based on complex system theory. Similarly, once the policy is implemented, the influence and effect of the same can be quantitatively figured by exploiting various data-driven techniques such as network analysis.

It was also established in a model engine that there can be unforeseen transitions that will purely depend on the rate of change of parameters. If the parameters of a system are changed in a rapid fashion, this change might lead to undesirable consequences. Society may respond very differently to a state policy if it is implemented in a short period of time. It is to be noted that the dynamics of a large change in policy implemented in a short period of time will be completely different from the policy changes that take years.

The Center for Computational Engineering and Networking, School of Engineering Coimbatore, Amrita Vishwa Vidyapeetham, Coimbatore campus, is involved in developing various data-driven tools and techniques to tackle a plethora of problems of national interest such as social media text analytics, stock market analysis, data analytics for point of care disease diagnosis and predictive techniques for smart grid stability, etc..

Share this Story: