Deep Learning

Flower in the crannied wall,
I pluck you out of the crannies,
I hold you here, root and all, in my hand,
Little flower—but if I could understand
What you are, root and all, and all in all,
I should know what God and man is.

Alfred Lord Tennyson

Artificial intelligence has caught the imagination of the public. Many dream of a world in the future where we will have autonomous cars, ubiquitous presence of robots and bots with whom one could not just have meaningful conversation but fulfilling relationship. While others are worried and concerned about the impact of machine learning and artificial intelligence on jobs and the future of work. There are also others worried about a monolithic artificial intelligence taking over the world, with disastrous consequences to humanity itself.

Some of the key areas of focus in the artificial intelligence space are autonomous driving, natural language processing and image recognition.  There are many questions that are still unanswered. There are still many questions that have not yet been asked.

How safe the automonous cars need to be, before they are considered safe enough? In safety don’t extreme outliers matters as much as averages? What would be the ethical consequences of an accident? How do we factor morality into a series of rules to be learnt by a machine? Does conversational fluency with a natural language correlate with knowledge and wisdom? Does language impact the way we think and as a corollary the way our minds are wired or is it the other way around, that we use language the way we think? How do we embed deeper constructs like empathy and compassion into natural language processing?

I had a chance to attend an inspiring lecture by Yoshua Bengio, co-author of the recent book on Deep Learning (an MIT Press book), made available online for free in the spirt of Open AI. My key take away from the lecture was though we have made very significant progress in several areas right from image recognition, natural language processing and autonomous driving both in academic and industry research groups, there is still a long way to go, before we can claim that we have come close to proclaming that we have truly understood the human brain all in all.