keerthana chelladurai

about everything and nothing

[keerthanachelladurai]

Clongriffin to Dun Laoghaire

The journey for the past 60 days where the destination always seems to excite me. Doing the things I love as this is officially my first job right after graduation. The feeling is still the same as Day 1, with everyday is somehow different from the previous ones.

One day it's just me and the trees running backwards or its me and the detective Hunter trying to catch the killer or its me and ARR or its just me aimlessly looking at the slowly passing beautiful scene out the dart window.

Some are chatting others are glued to their phones. Only some, very rarely reads. People here mind their business which I really appreciate as it was never the case back home. Nobody cares where you work or if you're married or if you're a sociopath.

Breathing in each moment and living life at my own terms. Hope this is forever and grateful for life as it is.

[keerthanachelladurai]

How Cooking Made Us Human

Soundcloud link: https://soundcloud.com/bridging-the-gaps/how-cooking-made-us-human-with-richard-wrangham?si=3a5317f83e864bcb811a780f6d1f56df&utm_source=clipboard&utm_medium=text&utm_campaign=social_sharing

The discussion starts with the Cooking Hypothesis: While studying chimpanzees, our guest Dr. Prof. Richard Wrangham only ate what the chimpanzees ate and discovered that he was not able to survive on that as the digestive system couldn’t able to adapt to it. The cooking hypothesis states that cooked food is better than wild rouge food for humans.

And the discussion went on about the age when humans started cooking their food which roughly was about 2500 years ago when they started to use fire. The main evolution of humans was from eating raw food to cooking our meals, made easier to digest, and gave higher calories. As the same changed the shape of the human which increased the brain size and decreased the size of the gut.

Two important objectives of cooking food should contain high digestibility and high energy content, the discussion went on about how this was not the case 30 years ago. And how cooking plants/ vegetables differ from animal meat, the different effects in different food, and the effect of the body breaking down the chunks to generate energy.

Consuming raw food depends on the environment you live in and how you are. Raw food has high protein but is certainly bad for infants and people who tend to stay in healthy shape. Why calories are not a good measure to calculate energy is because it's unrealistic and food varies in the amount of work we do as the calories value are less once it's cooked. Just like traffic lights, the author suggests we need to have some labels to calculate the calorie value.

Experiments about feeding cooked foods to animals to understand their digestion were just like humans when given a choice of choosing cooked vs raw food, most animals prefer their food cooked. Further studies have proved that increases digestion and reduces work of the body to digest it. The same was tested on snakes where the measure of Co2 and oxygen was used after they had a meal, where 25% of calories increase with cooked meat and reduced the body's work on digestion.

And the discussion ended with what makes us human? The ability to cook our food which no other species can! Cooked food -> easy digestion -> more energy.

The reason why I chose this topic is I just started cooking for the past eight months, before that I never ever tried it and didn't have any interest but since Necessity is the mother of invention I gradually started and I even kind of liked it. The topic caught my eye and my first thought was I've been human for only eight months!

Things that I learnt:

Cooking our food made it easy on digestion just like pre-processing the data made finding insights easier.

The author states that there are different effects on different types of food depending on physical and chemical state, as a data scientist I can relate it to different types of data that comes in from different data source and how each is unique and how we use different tools to prepare and process.

How raw food may have high protein but is not good for health as same as how raw data has hidden information that doesn’t make any sense with cleaning and pre-processing.

About the Guest: Dr. Richard Wrangham

More information about the guest can be found here: https://www.discoverlife.org/who/CV/Wrangham,_Richard.html

Dr. Prof. Richard Wrangham – an anthropologist and primatologist, born in Leeds [1] was a zoology undergrad at Oxford University and was a research assistant, got inspired from watching chimpanzee behavior and got his Ph.D. on the same and studying about it till date. He is well known for co-directing Kibale Chimpanzee Project in 1987 [2] the main focus of the project is based on the behavioral ecology of apes. This blog is about his book Catching Fire: How Cooking Made Us Human.

Reference:

[1] Wikipedia https://en.wikipedia.org/wiki/Richard_Wrangham

[2] Kibale Chimpanzee Project: https://kibalechimpanzees.wordpress.com/2013/09/05/mission/  

Mission

The Kibale Chimpanzee Project is dedicated to the conservation and welfare of chimpanzees and their habitats. We are committed to promoting

[keerthanachelladurai]

Asking Better Questions for Problem Solving, Innovation and Effective Leadership with Hal Gregersen

Soundcloud link: https://soundcloud.com/bridging-the-gaps/asking-better-questions-for-problem-solving-innovation-and-effective-leadership-hal-gregersen?si=c7e0efa918d94d839f70558e8cdb66ed&utm_source=clipboard&utm_medium=text&utm_campaign=social_sharing

The first question in the podcast was about how the guest Hal Gregersen got into research in the field of leadership, the interesting answer for that was his parents as his mother has a keen observation of the surroundings. The discussion starts on asking the right questions which led to the guest getting into grad school and got the Ph.D. in leadership. The interesting statement by Hal that I personally agree was you can lead transformation and change unless you ask right questions. Host led the discussion about evolution of leadership landscape over the years and currently about digital transformation leadership, our guest gave an exemplary example on how Mary Parker operated in men’s working world and how she become the exception women in her field, how she led a group of people on discussing meeting years back to get solutions for the right questions. This happened hundred years ago, today in the digital world with so many things happening with machine learning, deep learning we are in the path of change.

Quoted by the guest “leaders create a psychological space where inquiry leads to insight, insight – we learnt something we didn’t know before, and leads to positive impact”. In a nutshell, questions -> insight -> knowledge.

Discussion went on about catalytic questions, which are the ultimate answers. It reframes the fundamentally off assumption but generated the energy to make it work. The discussion then went on about how to build amental model, if and only if we have the courage to ask more question and the criticism doesn’t affect you. My view in this is, some are embarrassed to ask questions thinking that they may look silly or thinks about what other people will think about them, because of that they lose some great opportunities, when one dare to ask question, he rise above and beyond.

Good questions are recursive, asking series of whys leads to modify and create better systems keeping in mind what we started in the beginning in the sense without deviating from the use case. Discussion then went on about how sustainable creativity is not possible without any vision for industries, we can create a better innovative world by being creative.

Things that I learned:

Asking better questions keeps the system (both humans and machines) without underfitting or overfitting but keeps it neutralized. Going on the right path and recognizing the patterns are more important than changing the course of track which might lead to not having the right answers.

Mind map of questions – knowing the relationship of questions. History of asking questions and maintaining ledger-type notes on questions that you think of, wanting to ask, curious enough to know the answers at the end of the day take a step back and analyze the questions. Our guest calls this question audits. And I think this will help everyone on the path of the betterment of one’s knowledge. Ask gazillions of questions whenever you’re stuck.

About the Guest: Hal Gregersen (https://halgregersen.com/)

Storyteller, researcher, questioner, photographer and explorer. Author of Questions are the answers and other exceptional books is a senior lecturer in leadership and innovation at MIT's Sloan School of Management, a former executive director of the MIT Leadership Center, a fellow at Innosight, and a co-founder of the Innovator’s DNA consulting group. He has been Ranked one of the world’s 20 most influential management thinkers by Thinkers50, and the winner of the 2017 Distinguished Achievement Award for leadership. Hal’s publications can be found here (https://mitsloan.mit.edu/faculty/directory/hal-gregersen). He has also authored n number of articles in Forbes (https://www.forbes.com/sites/halgregersen/?sh=1ccccde740de) about leadership and innovative growth.  

[keerthanachelladurai]

Machines like Us: TOWARD AI WITH COMMON SENSE

Soundcloud link: https://soundcloud.com/bridging-the-gaps/machines-like-us-toward-ai-with-common-sense-with-professor-ronald-brachman?si=ea1cef08fb4c4c61bd9a90670a789eb8&utm_source=clipboard&utm_medium=text&utm_campaign=social_sharing

The discussion starts about how the guest Prof. Ronal Brachman sees the growth of AI but in a narrow way, in the sense performing singular tasks and not in a general way. The discussion went on about how AI can play chess and checkers. Because the computation power in those days, the tasks were aimed at doing single tasks, those were the starting point of systems that we have now to name a few, self-driving cars, personal assistance and so on. And the conversation started to dig deep on computer vision, deep learning and the data available to train these complex algorithms on important tasks.

Move from general AI approach – ultimate goal is to develop self-controlling AI and to develop human level intelligence. Interesting example given by the guest was sending robot to mars (mars rover). With the current robots that can’t be re-tasked when found something interesting like a metal in mars. More discussion on autonomous AI, where there is a level of knowledge and reasoning for the actions that carried out by the AI. Example about autonomous cars on making decisions on the go was discussed.

Failures on AI: Alexa gave a challenge to 10-year-old to plugging a phone partially and insert a penny between the probes, which suggest the lack of common sense or failure of AI. And another example was discussed on AI bot for suicide prevention, where the bot encouraged the person to go ahead with the suicide thought.

More insight on common sense in humans, where we remember or forget things, but comes from experience. And common sense on machines comes from the same, as an example when a cake was hidden in a washing machine for surprising someone, the AI should be intelligent enough not to turn it on, which was a great example!

The conversation went on about how in 30 to 40 years experts say that we will have fully generalized AI, our guest's thought is that it might be not possible but will take more time to get there. The discussion ended on McCarthy's programs with common sense, the ability to take responsibility and decisions outside the guidance, the advice taker.

Things that I learnt from the podcast:

Developing AI that is more autonomous and logical. Can make own decisions that don't cause any harm around.

Reinforcement learning on AI, training on the fly and experience where to take responsibility of their actions

Dangers on AI making mistakes that humans can’t predict, which could be destructible in a massive way. Which taught me as a data scientist that I should be more careful on the systems that I will be developing.

About the Guest: Professor Ronald Brachman (http://www.brachman.org/mybackground.html)

Author of Machines like us – towards AI with common sense

Pioneer in the field of Artificial Intelligence for more than 40 years. he served as President of Association for the Advancement of Artificial Intelligence from 2003 – 2006.  Before that he was the Secretary-Treasurer of International Joint Conferences on Artificial Intelligence, Inc for 9 years. He is currently co-editing the Synthesis lecture series on AI and ML. He was also a member and Treasurer of the Board of Directors of the Computing Research Association. He was an advisor for start-up called Segovia from 2015 – 2019.

He was the Chief Scientist of Yahoo and the Head of Yahoo Labs from 2012 - 1016.  Prior to that, starting in September of 2005, I was the Associate Head of the Labs and essentially its Chief Operating Officer.  before joining Yahoo in 2005, he was the Director of the Information Processing Technology Office at DARPA, the Defense Advanced Research Projects Agency.  The Personalized Assistant that Learns program that he created with help from Zach Lemnios led directly to the technology that Apple now offers in the form of Siri on its iPhone products.

More interesting articles by Prof Ronald can be found here https://scholar.google.com/citations?user=zMnT8BsAAAAJ&hl=en