Profile cover photo
Profile photo
Vijay Chidambaram
Vijay's posts

Post has shared content
This is one of those ideas which is superficially good and becomes terrible if you look at it more closely. The concept is that in sentencing, we should do a risk assessment, based on objective criteria rather than on the prejudices of the individuals in the court, and use that to influence the decision. Those more likely to re-offend, in this system, would receive harsher and longer sentences.

The first problem with this is that if you build a 100% honest model of re-offending probability, what you're building is a model of your own system, not of the person. For example: If you lock someone in prison for a few years, offering them no training or rehabilitation, and then upon release have various penalties on them which basically prevent them from getting a job -- ranging from the simple "nobody wants to hire someone with a conviction record" to "explicit legal bars to their getting certain kinds of job, living in certain areas, etc" -- then it will probably not surprise you that this person is significantly more likely to turn to a professional life of crime. A model which correctly recognized and predicted this would therefore conclude that the only solution is to lock this person up for life, since at any point after they're released, they're simply likely to become criminals again.

This highlights the deeper problem in such a model, of course, which is that its basic design parameters, where the only variable it controls is "imprison more" or "imprison less," create a false dichotomy: rather than asking "which course of action is most likely to lead to the person no longer engaging in crime," it only considers one possible course of action, and that action (again, by the design of the system) most often increases the probability of future crime. 

The criticism of this system that it will end up encoding implicit racial biases is only sort-of correct. This model will definitely end up having a strong racial component; even if you eliminate race as an input, your race is so strongly correlated to other things like where you live that the system will end up modeling your race, and basing its decisions upon that, one way or the other. And that will, indeed, end up increasing sentences for Black and Latino offenders, for all the reasons specified above.

But in this case, the racial biases which the system would acquire are simply one manifestation of the even deeper and more profound problem that this model is simply designed to optimize for throwing people into prison.

If you want a variation on this which actually works, give the model access to a wide range of possible consequences, and ask it which of those will minimize the odds of re-offense, presumably balanced against various costs. You'll almost certainly find that rehabilitation, training, and treatment overwhelmingly work best to minimize that. (And in the cases where they don't, your best bet is likely to simply take them out and shoot them)

I would actually quite strongly favor such a project, because it would require its creators to make very explicit the thing for which they are trying to optimize. You can't lie to a computer about what you want it to do; if you want to minimize the chance of re-offense, you have to tell it to do so. If you instead want to optimize the system for retribution, or to cow a broad population into submission, or to maximize revenue, the model will absolutely be able to do that as well -- but you would have to tell it explicitly to do so, and it's very hard to lie to yourself about that.

h/t +Amy Quispe over on Twitter for prompting me to actually write about this one.

Post has shared content
Bimal Roy was, until a few days ago, the director of the Indian Statistical Institute, a graduate institute based in Kolkata. He is well-known in cryptography for his work on block ciphers. He was also awarded the Padma Shri last year, one of the highest civilian awards in India. 

My interactions with him were limited, but I have generally been impressed by the huge summer internship program for Indian undergraduates that he runs at ISI. 

Roy was just removed from his position by the Ministry of Statistics and Programme Implementation, citing concerns about potential "indiscipline". The reasons are almost comically vague:

"A number of general and specific matters of financial and administrative irregularities which show the direct or supervisory responsibilities for acts of omission or commission on the part of the present Director, Prof. B. Roy are available in the Ministry in the various files on the different subjects. "

Below is an account of the events leading up to his dismissal written by Roy's colleague and former Ph.D. student, Sushmita Ruj. I copy her email here verbatim.

I don't know the facts of the story, but I am inclined to believe Sushmita's account. The lack of transparency should be especially troubling for academics, in India and elsewhere. 


This is Sushmita Ruj, writing from the Cryptology group of Indian Statistical Institute, Kolkata.

Professor Bimal Roy, the leader of our group, has been removed from the post of Director of the Indian Statistical Institute with a special emergency order issued by out governing Ministry (MoS&PI) on 10 June 2015. He was about to complete his five year term as Director on 31 July 2015.

The Ministry has expressed its apprehension that Prof. Bimal Roy may indulge in "propagation of indiscipline and mischief, including acts of financial and administrative impropriety" if he continues as the Director. However, neither has it put forward any formal charge against Prof. Roy, nor has it offered him the basic courtesy to defend himself before the decision was taken. There was no show-cause served to Prof. Roy at all, and the grounds for the decision seem quite fragile:

As we understand, this drastic decision from the Ministry is orchestrated by the Chairman of our Institute, Mr. Arun Shourie, who allegedly tried to coerce Prof. Bimal Roy into signing the proceedings of our last Council Meeting, in which the selection of the new Director of the Institute was considered. Allegedly, the drafted proceedings of the meeting was substantially different from the actual incidents at the meeting, especially on the issue of the selection of the new Director, and hence Prof. Roy refused to sign the proceedings. This is the only issue that may have been considered as "indiscipline" on Prof. Roy's part, as far as we know.

To verify the truth in this matter, a senior Professor of our Institute filed a written petition (RTI) to obtain the audio recording of the meeting, and the hearing of the RTI was due on 11 June 2015. It is quite an amazing coincidence that the Ministry took the decision of stripping Prof. Bimal Roy of all his powers as the Director just the night before, in the afternoon of 10 June 2015. Not to mention, the RTI hearing has been postponed at once. This raises a strong suspicion that the Ministry, supporting the Chairman of the Institute, is aiming at suppressing the audio recording of the meeting to avoid a transparent investigation into the matters at hand.

Professor Bimal Roy has dedicated his life for services to the Nation, and this is not what he deserves from the Institute which has gained from his leadership in the last five years more than it probably has under the leadership of any other Director in the recent times. Prof. Roy, and our Institute, deserves a transparent independent public investigation. The audio recording and proceedings of the Council meeting should be released and investigated by independent authorities. Prof. Roy should be offered a chance for a fair trial, independent of the actions by the Ministry, and till the investigation is over, Prof. Roy should be reinstated as the Director of our Institute, till 31 July 2015, the end of his rightful term. At the end of his term, Prof. Bimal Roy deserves a vote of thanks from the Council for his services to the Institute during the last five years, and not a humiliating send-off. This is simply unacceptable!

If you support the cause, and want justice for Prof. Bimal Roy, please convey your thoughts to the Hon'ble President of India at the following email IDs.     
Public Grievance Cell : Mr. Purushottam Dass   
Private Secretary to the President : Mr. Rajneesh   
Private Secretary to the President : Mr. Pradeep Gupta

You may also write directly to the Hon'ble Prime Minister of India through the online portal at :

I request you to propagate the news in the concerned academic community, government organizations and public media that you deem suitable, and sign-and-share the online petition supporting the cause.

It will be nice to have the community support Prof. Bimal Roy in these dire times. Please feel free to write to him directly as well.


Post has shared content

Post has shared content
" Some of you, not all of you, are working right now on another app for people to share pictures of food or a social network for dogs. I am here to tell you that your country has a better use for your talents."

I love +Mikey D​. Read this. Consider it carefully. 

Post has shared content

Post has shared content

Post has attachment

Post has shared content

Post has shared content

Post has shared content
+Eric Price blogs about the NIPS experiment on +Moritz Hardt's blog: 

Summary: Half the papers appearing at NIPS would be rejected if the review process was repeated. 
Wait while more posts are being loaded