For you who are worried about the threat of artificial intelligence, MIRI seems to make it
their main concern. Look up their website and subscribe. On my list of existential
threats it comes well below natural stupidity.
Brent
-------- Original Message --------
Subject: The Machine Intelligence Research Institute Blog
Date: Fri, 05 Sep 2014 12:07:00 +0000
From: Machine Intelligence Research Institute » Blog <[email protected]>
To: [email protected]
Machine Intelligence Research Institute » Blog
The Machine Intelligence Research Institute Blog <http://intelligence.org>
------------------------------------------------------------------------------------------
John Fox on AI safety
<http://intelligence.org/2014/09/04/john-fox/?utm_source=rss&utm_medium=rss&utm_campaign=john-fox>
Posted: 04 Sep 2014 12:00 PM PDT
John Fox portrait John Fox <http://www.cossac.org/people/johnfox> is an interdisciplinary
scientist with theoretical interests in AI and computer science, and an applied focus in
medicine and medical software engineering. After training in experimental psychology at
Durham and Cambridge Universities and post-doctoral fellowships at CMU and Cornell in the
USA and UK (MRC) he joined the Imperial Cancer Research Fund (now Cancer Research UK) in
1981 as a researcher in medical AI. The group’s research was explicitly multidisciplinary
and it subsequently made significant contributions in basic computer science, AI and
medical informatics, and developed a number of successful technologies which have been
commercialised.
In 1996 he and his team were awarded the 20th Anniversary Gold Medal of the European
Federation of Medical Informatics for the development of PROforma, arguably the first
formal computer language for modeling clinical decision and processes. Fox has published
widely in computer science, cognitive science and biomedical engineering, and was the
founding editor of the /Knowledge Engineering Review / (Cambridge University Press).
Recent publications include a research monograph /Safe and Sound: Artificial Intelligence
in Hazardous Applications
<http://smile.amazon.com/Safe-Sound-Artificial-Intelligence-Applications/dp/0262062119/ref=nosim?tag=793775876-20>/
(MIT Press, 2000) which deals with the use of AI in safety-critical fields such as medicine.
*Luke Muehlhauser*: You’ve spent many years studying AI safety issues, in particular in
medical contexts, e.g. in your 2000 book with Subrata Das, /Safe and Sound: Artificial
Intelligence in Hazardous Applications
<http://smile.amazon.com/Safe-Sound-Artificial-Intelligence-Applications/dp/0262062119/ref=nosim?tag=793775876-20>/.
What kinds of AI safety challenges have you focused on in the past decade or so?
------------------------------------------------------------------------------------------
*John Fox*: From my first research job, as a post-doc with AI founders Allen Newell and
Herb Simon at CMU, I have been interested in computational theories of high level
cognition. As a cognitive scientist I have been interested in theories that subsume a
range of cognitive functions, from perception and reasoning to the uses of knowledge in
autonomous decision-making. After I came back to the UK in 1975 I began to combine my
theoretical interests with the practical goals of designing and deploying AI systems in
medicine.
Since our book was published in 2000 I have been committed to testing the ideas in it by
designing and deploying many kind of clinical systems, and demonstrating that AI
techniques can significantly improve quality and safety of clinical decision-making and
process management. Patient safety is fundamental to clinical practice so, alongside the
goals of building systems that can improve on human performance, safety and ethics have
always been near the top of my research agenda.
------------------------------------------------------------------------------------------
*Luke Muehlhauser:* Was it straightforward to address issues like safety and ethics in
practice?
------------------------------------------------------------------------------------------
*John Fox*: While our concepts and technologies have proved to be clinically successful we
have not achieved everything we hoped for. Our attempts to ensure, for example, that
practical and commercial deployments of AI technologies should explicitly honor ethical
principles and carry out active safety management have not yet achieved the traction that
we need to achieve. I regard this as a serious cause for concern, and unfinished business
in both scientific and engineering terms.
The next generation of large-scale knowledge based systems and software agents that we are
now working on will be more intelligent and will have far more autonomous capabilities
than current systems. The challenges for human safety and ethical use of AI that this
implies are beginning to mirror those raised by the singularity hypothesis. We have much
to learn from singularity researchers, and perhaps our experience in deploying autonomous
agents in human healthcare will offer opportunities to ground some of the singularity
debates as well.
------------------------------------------------------------------------------------------
*Luke*: You write that your “attempts to ensure… [that] commercial deployments of AI
technologies should… carry out active safety management” have not yet received as much
traction as you would like. Could you go into more detail on that? What did you try to
accomplish on this front that didn’t get adopted by others, or wasn’t implemented?
------------------------------------------------------------------------------------------
*John*: Having worked in medical AI from the early ‘seventies I have always been keenly
aware that while AI can help to mitigate the effects of human error there is a potential
downside too. AI systems could be programmed incorrectly, or their knowledge could
prescribe inappropriate practices, or they could have the effect of deskilling the human
professionals who have the final responsibility for their patients. Despite well-known
limitations of human cognition people remain far and away the most versatile and creative
problem solvers on the planet.
In the early ‘nineties I had the opportunity to set up a project whose goal was to
establish a rigorous framework for the design and implementation of AI systems for safety
critical applications. Medicine was our practical focus but the RED project^1
<http://intelligence.org/2014/09/04/john-fox/#footnote_0_11319> was aimed at the
development of a general architecture for the design of autonomous agents that could be
trusted to make decisions and carry out plans as reliably and safely as possible,
certainly to be as competent and hence as trustworthy as human agents in comparable tasks.
This is obviously a hard problem but we made sufficient progress
<http://www.tandfonline.com/doi/abs/10.1080/095281397146979> on theoretical issues and
design principles that I thought there was a good chance the techniques might be
applicable in medical AI and maybe even more widely.
I thought AI was like medicine, where we all take it for granted that medical equipment
and drug companies have a duty of care to show that their products are effective and safe
before they can be certificated for commercial use. I also assumed that AI researchers
would similarly recognize that we have a “duty of care” to all those potentially affected
by poor engineering or misuse in safety critical settings but this was naïve. The
commercial tools that have been based on the technologies derived from AI research have to
date focused on just getting and keeping customers and safety always takes a back seat.
In retrospect I should have predicted that making sure that AI products are safe is not
going to capture the enthusiasm of commercial suppliers. If you compare AI apps with drugs
we all know that pharmaceutical companies have to be firmly regulated to make sure they
fulfill their duty of care to their customers and patients. However proving drugs are safe
is expensive and also runs the risk of revealing that your new wonder-drug isn’t even as
effective as you claim! It’s the same with AI.
I continue to be surprised how optimistic software developers are – they always seem to
have supreme confidence that worst-case scenarios wont happen, or that if they do happen
then their management is someone else’s responsibility. That kind of technical
over-confidence has led to countless catastrophes in the past, and it amazes me that it
persists.
There is another piece to this, which concerns the roles and responsibilities of AI
researchers. How many of us take the risks of AI seriously so that it forms a part of our
day-to-day theoretical musings and influences our projects? MIRI has put one worst case
scenario in front of us – the possibility that our creations might one day decide to
obliterate us – but so far as I can tell the majority of working AI professionals either
see safety issues as irrelevant to the pursuit of interesting scientific questions or,
like the wider public, that the issues are just science fiction.
I think experience in medical AI trying to articulate and cope with human risk and safety
may have a couple of important lessons for the wider AI community. First we have a duty of
care that professional scientists cannot responsibly ignore. Second, the AI business will
probably need to be regulated, in much the same way as the pharmaceutical business is. If
these propositions are correct then the AI research community would be wise to engage with
and lead on discussions around safety issues if it wants to ensure that the regulatory
framework that we get is to our liking!
------------------------------------------------------------------------------------------
*Luke*: Now you write, “That kind of technical over-confidence has led to countless
catastrophes in the past…” What are some example “catastrophes” you’re thinking of?
------------------------------------------------------------------------------------------
*John*:
Psychologists have known for years that human decision-making is flawed, even if amazingly
creative sometimes, and overconfidence is an important source of error in routine
settings. A large part of the motivation for applying AI in medicine comes from the
knowledge that, in the words of the Institute of Medicine, “To err is human” and
overconfidence is an established cause of clinical mistakes.^2
<http://intelligence.org/2014/09/04/john-fox/#footnote_1_11319>
Over-confidence and its many relatives (complacency, optimism, arrogance and the like)
have a huge influence on our personal successes and failures, and our collective futures.
The outcomes of the US and UK’s recent adventures around the world can be easily
identified as consequences of overconfidence, and it seems to me that the polarized
positions about global warming and planetary catastrophe are both expressions of
overconfidence, just in opposite directions.
------------------------------------------------------------------------------------------
*Luke*: Looking much further out… if one day we can engineer AGIs
<http://intelligence.org/2013/08/11/what-is-agi/>, do you think we are likely to figure
out how to make them safe?
------------------------------------------------------------------------------------------
*John*: History says that making any technology safe is not an easy business. It took
quite a few boiler explosions before high-pressure steam engines got their iconic
centrifugal governors. Ensuring that new medical treatments are safe as well as effective
is famously difficult and expensive. I think we should assume that getting to the point
where an AGI manufacturer could guarantee its products are safe will be a hard road, and
it is possible that guarantees are not possible in principle. We are not even clear yet
what it means to be “safe”, at least not in computational terms.
It seems pretty obvious that entry level robotic products like the robots that carry out
simple domestic chores or the “nursebots” that are being trialed for hospital use, have
such a simple repertoire of behaviors that it should not be difficult to design their
software controllers to operate safely in most conceivable circumstances. Standard safety
engineering techniques like HAZOP^3
<http://intelligence.org/2014/09/04/john-fox/#footnote_2_11319> are probably up to the job
I think, and where software failures simply cannot be tolerated software engineering
techniques like formal specification and model-checking are available.
There is also quite a lot of optimism around more challenging robotic applications like
autonomous vehicles and medical robotics. Moustris et al.^4
<http://intelligence.org/2014/09/04/john-fox/#footnote_3_11319> say that autonomous
surgical robots are emerging that can be used in various roles, automating important steps
in complex operations like open-heart surgery for example, and they expect them to become
standard in – and to revolutionize the practice of – surgery. However at this point it
doesn’t seem to me that surgical robots with a significant cognitive repertoire are
feasible and a human surgeon will be in the loop for the foreseeable future.
------------------------------------------------------------------------------------------
*Luke*: So what might artificial intelligence learn from natural intelligence?
------------------------------------------------------------------------------------------
As a cognitive scientist working in medicine my interests are co-extensive with those of
scientists working on AGIs. Medicine is such a vast domain that practicing it safely
requires the ability to deal with countless clinical scenarios and interactions and even
when working in a single specialist subfield requires substantial knowledge from other
subfields. So much so that it is now well known that even very experienced humans with a
large clinical repertoire are subject to significant levels of error.^5
<http://intelligence.org/2014/09/04/john-fox/#footnote_4_11319> An artificial intelligence
that could be helpful across medicine will require great versatility, and this will
require a general understanding of medical expertise and a range of cognitive capabilities
like reasoning, decision-making, planning, communication, reflection, learning and so forth.
If human experts are not safe is it well possible to ensure that an AGI, however
sophisticated, will be? I think that it is pretty clear that the range of techniques
currently available for assuring system safety will be useful in making specialist AI
systems reliable and minimizing the likelihood of errors in situations that their human
designers can anticipate. However, AI systems with general intelligence will be expected
to address scenarios and hazards that are beyond us to solve currently and often beyond
designers even to anticipate. I am optimistic but at the moment I don’t see any convincing
reason to believe that we have the techniques that would be sufficient to guarantee that a
clinical super-intelligence is safe, let alone an AGI that might be deployed in many domains.
------------------------------------------------------------------------------------------
*Luke*: Thanks, John!
------------------------------------------------------------------------------------------
1. Rigorously Engineered Decisions <http://www.cossac.org/projects/red>
2. Overconfidence in major disasters:
• D. Lucas. /Understanding the Human Factor in Disasters./
<http://www.maneyonline.com/doi/abs/10.1179/isr.1992.17.2.185>
Interdisciplinary
Science Reviews. Volume 17 Issue 2 (01 June 1992), pp. 185-190.
• “Nuclear safety and security.
<http://en.wikipedia.org/wiki/Nuclear_safety_and_security>
Psychology of overconfidence:
• Overconfidence effect. <http://en.wikipedia.org/wiki/Overconfidence_effect>
• C. Riordan. Three Ways Overconfidence Can Make a Fool of You
<http://www.forbes.com/sites/forbesleadershipforum/2013/01/08/three-ways-overconfidence-can-make-a-fool-of-you/>
Forbes Leadership Forum.
Overconfidence in medicine:
• R. Hanson. Overconfidence Erases Doc Advantage.
<http://www.overcomingbias.com/2007/04/overconfidence_.html> Overcoming
Bias, 2007.
• E. Berner, M. Graber. Overconfidence as a Cause of Diagnostic Error in
Medicine.
<http://www.amjmed.com/article/S0002-9343%2808%2900040-5/abstract> The
American
Journal of Medicine. Volume 121, Issue 5, Supplement, Pages S2–S23, May 2008.
• T. Ackerman. Doctors overconfident, study finds, even in hardest cases.
<http://www.houstonchronicle.com/news/health/article/Doctors-overconfident-study-finds-even-in-4766096.php>
Houston Chronicle, 2013.
General technology example:
• J. Vetter, A. Benlian, T. Hess. Overconfidence in IT Investment Decisions:
Why
Knowledge can be a Boon and Bane at the same Time.
<http://aisel.aisnet.org/icis2011/proceedings/generaltopics/4/> ICIS 2011
Proceedings.
Paper 4. December 6, 2011.
3. Hazard and operability study
<http://en.wikipedia.org/wiki/Hazard_and_operability_study>
4. Int J Med Robotics Comput Assist Surg 2011; 7: 375–39
5. A. Ford. Domestic Robotics – Leave it to Roll-Oh, our Fun loving Retrobot
<http://ieet.org/index.php/IEET/more/ford20140702>. Institute for Ethics and
Emerging
Technologies, 2014.
The post John Fox on AI safety <http://intelligence.org/2014/09/04/john-fox/> appeared
first on Machine Intelligence Research Institute <http://intelligence.org>.
Daniel Roy on probabilistic programming and AI
<http://intelligence.org/2014/09/04/daniel-roy/?utm_source=rss&utm_medium=rss&utm_campaign=daniel-roy>
Posted: 04 Sep 2014 08:03 AM PDT
Daniel Roy portrait Daniel Roy <http://danroy.org/> is an Assistant Professor of
Statistics at the University of Toronto. Roy earned an S.B. and M.Eng. in Electrical
Engineering and Computer Science, and a Ph.D. in Computer Science, from MIT. His
dissertation on probabilistic programming received the department’s George M Sprowls
Thesis Award. Subsequently, he held a Newton International Fellowship of the Royal
Society, hosted by the Machine Learning Group at the University of Cambridge, and then
held a Research Fellowship at Emmanuel College. Roy’s research focuses on theoretical
questions that mix computer science, statistics, and probability.
*Luke Muehlhauser*: The abstract of Ackerman, Freer, and Roy (2010)
<http://danroy.org/papers/AckFreRoy-CompCondProb-preprint.pdf> begins:
As inductive inference and machine learning methods in computer science see
continued
success, researchers are aiming to describe even more complex probabilistic
models and
inference algorithms. What are the limits of mechanizing probabilistic
inference? We
investigate the computability of conditional probability… and show that
there are
computable joint distributions with noncomputable conditional distributions,
ruling
out the prospect of general inference algorithms.
In what sense does your result (with Ackerman & Freer) rule out the prospect of general
inference algorithms?
------------------------------------------------------------------------------------------
*Daniel Roy*: First, it’s important to highlight that when we say “probabilistic
inference” we are referring to the problem of computing conditional probabilities
<http://en.wikipedia.org/wiki/Conditional_probability>, while highlighting the role of
conditioning in Bayesian statistical analysis.
Bayesian inference centers around so-called posterior distributions. From a subjectivist
standpoint, the posterior represents one’s updated beliefs after seeing (i.e.,
conditioning on) the data. Mathematically, a posterior distribution is simply a
conditional distribution (and every conditional distribution can be interpreted as a
posterior distribution in some statistical model), and so our study of the computability
of conditioning also bears on the problem of computing posterior distributions, which is
arguably one of the core computational problems in Bayesian analyses.
Second, it’s important to clarify what we mean by “general inference”. In machine learning
and artificial intelligence (AI), there is a long tradition of defining formal languages
in which one can specify probabilistic models over a collection of variables. Defining
distributions can be difficult, but these languages can make it much more straightforward.
The goal is then to design algorithms that can use these representations to support
important operations, like computing conditional distributions. Bayesian networks can be
thought of as such a language: You specify a distribution over a collection of variables
by specifying a graph over these variables, which breaks down the entire distribution into
“local” conditional distributions corresponding with each node, which are themselves often
represented as tables of probabilities (at least in the case where all variables take on
only a finite set of values). Together, the graph and the local conditional distributions
determine a unique distribution over all the variables.
An inference algorithms that support the entire class of all finite, discrete, Bayesian
networks might be called general, but as a class of distributions, those having finite,
discrete Bayesian networks is a rather small one.
In this work, we are interested in the prospect of algorithms that work on very large
classes of distributions. Namely, we are considering the class of samplable distributions,
i.e., the class of distributions for which there exists a probabilistic program that can
generate a sample using, e.g., uniformly distributed random numbers or independent coin
flips as a source of randomness. The class of samplable distributions is a natural one:
indeed it is equivalent to the class of computable distributions, i.e., those for which we
can devise algorithms to compute lower bounds on probabilities from descriptions of open
sets. The class of samplable distributions is also equivalent to the class of
distributions for which we can compute expectations from descriptions of bounded
continuous functions.
The class of samplable distributions is, in a sense, the richest class you might hope to
deal with. The question we asked was: is there an algorithm that, given a samplable
distribution on two variables X and Y, represented by a program that samples values for
both variables, can compute the conditional distribution of, say, Y given X=x, for almost
all values for X? When X takes values in a finite, discrete set, e.g., when X is binary
valued, there is a general algorithm, although it is inefficient. But when X is
continuous, e.g., when it can take on every value in the unit interval [0,1], then
problems can arise. In particular, there exists a distribution on a pair of numbers in
[0,1] from which one can generate perfect samples, but for which it is impossible to
compute conditional probabilities for one of the variables given the other. As one might
expect, the proof reduces the halting problem to that of conditioning a specially crafted
distribution.
This pathological distribution rules out the possibility of a general algorithm for
conditioning (equivalently, for probabilistic inference). The paper ends by giving some
further conditions that, when present, allow one to devise general inference algorithms.
Those familiar with computing conditional distributions for finite-dimensional statistical
models will not be surprised that conditions necessary for Bayes’ theorem are one example.
------------------------------------------------------------------------------------------
*Luke*: In your dissertation (and perhaps elsewhere) you express a particular interest in
the relevance of probabilistic programming to AI, including the original aim of AI to
build machines which rival the general intelligence of a human. How would you describe the
relevance of probabilistic programming to the long-term dream of AI?
------------------------------------------------------------------------------------------
*Daniel*: If you look at early probabilistic programming systems, they were built by AI
researchers: De Raedt, Koller, McAllester, Muggleton, Pfeffer, Poole, Sato, to name a few.
The Church language, which was introduced in joint work with Bonawitz, Mansinghka,
Goodman, and Tenenbaum while I was a graduate student at MIT, was conceived inside a
cognitive science laboratory, foremost to give us a language rich enough to express the
range of models that people were inventing all around us. So, for me, there’s always been
a deep connection. On the other hand, the machine learning community as a whole is
somewhat allergic to AI and so the pitch to that community has more often been pragmatic:
these systems may someday allow experts to conceive, prototype, and deploy much larger
probabilistic systems, and at the same time, empower a much larger community of nonexperts
to use probabilistic modeling techniques to understand their data. This is the basis for
the DARPA PPAML <http://ppaml.galois.com/wiki/> program, which is funding 8 or so teams to
engineer scalable systems over the next 4 years.
From an AI perspective, probabilistic programs are an extremely general representation of
knowledge, and one that identifies uncertainty with stochastic computation. Freer,
Tenenbaum, and I recently wrote a book chapter
<http://danroy.org/papers/FreRoyTen-Turing.pdf> for the Turing centennial that uses a
classical medical diagnosis example to showcase the flexibility of probabilistic programs
and a general QUERY operator for performing probabilistic conditioning. Admittedly, the
book chapter ignores the computational complexity of the QUERY operator, and any serious
proposal towards AI cannot do this indefinitely. Understanding when we can hope to
efficiently update our knowledge in light of new observations is a rich source of research
questions, both applied and theoretical, spanning not only AI and machine learning, but
also statistics, probability, physics, theoretical computer science, etc.
------------------------------------------------------------------------------------------
*Luke*: Is it fair to think of QUERY as a “toy model” that we can work with in concrete
ways to gain more general insights into certain parts of the long-term AI research agenda,
even though QUERY is unlikely to be directly implemented in advanced AI systems? (E.g.
that’s how I think of AIXI.)
------------------------------------------------------------------------------------------
*Daniel*: I would hesitate to call QUERY a toy model. Conditional probability is a
difficult concept to master, but, for those adept at reasoning about the execution of
programs, QUERY demystifies the concept considerably. QUERY is an important conceptual
model of probabilistic conditioning.
That said, the simple guess-and-check algorithm we present in our Turing article runs in
time inversely proportional to the probability of the event/data on which one is
conditioning. In most statistical settings, the probability of a data set decays
exponentially towards 0 as a function of the number of data points, and so guess-and-check
is only useful for reasoning with toy data sets in these settings. It should come as no
surprise to hear that state-of-the-art probabilistic programming systems work nothing like
this.
On the other hand, QUERY, whether implemented in a rudimentary fashion or not, can be used
to represent and reason probabilistically about arbitrary computational processes, whether
they are models of the arrival time of spam, the spread of disease through networks, or
the light hitting our retinas. Computer scientists, especially those who might have had a
narrow view of the purview of probability and statistics, will see a much greater overlap
between these fields and their own once they understand QUERY.
To those familiar with AIXI, the difference is hopefully clear: QUERY performs
probabilistic reasoning in a model given as input. AIXI, on the other hand, is itself a
“universal” model that, although not computable, would likely predict (hyper)intelligent
behavior, were we (counterfactually) able to perform the requisite probabilistic
inferences (and feed it enough data). Hutter gives an algorithm implementing an
approximation to AIXI, but its computational complexity still scales exponentially in
space. AIXI is fascinating in many ways: If we ignore computational realities, we get a
complete proposal for AI. On the other hand, AIXI and its approximations take maximal
advantage of this computational leeway and are, therefore, ultimately unsatisfying. For
me, AIXI and related ideas highlight that AI must be as much a study of the particular as
it of the universal. Which potentially unverifiable, but useful, assumptions will enable
us to efficiently represent, update, and act upon knowledge under uncertainty?
------------------------------------------------------------------------------------------
*Luke*: You write that “AI must be as much a study of the particular as it is of the
universal.” Naturally, most AI scientists are working on the particular, the near term,
the applied. In your view, what are some other examples of work on the universal, in AI?
Schmidhuber’s Gödel machine comes to mind, and also some work that is as likely to be done
in a logic or formal philosophy department as a computer science department — e.g. perhaps
work on logical priors — but I’d love to hear what kinds of work you’re thinking of.
------------------------------------------------------------------------------------------
*Daniel*: I wouldn’t equate any two of the particular, near-term, or applied. By the word
particular, I am referring to, e.g., the way that our environment affects, but is also
affected by, our minds, especially through society. More concretely, both the physical
spaces in which most of us spend our days and the mental concepts we regularly use to
think about our daily activities are products of the human mind. But more importantly,
these physical and mental spaces are necessarily ones that are easily navigated by our
minds. The coevolution by which this interaction plays out is not well studied in the
context of AI. And to the extent that this cycle dominates, we would expect a universal AI
to be truly alien. On the other hand, exploiting the constraints of human constructs may
allow us to build more effective AIs.
As for the universal, I have an interest in the way that noise can render idealized
operations computable or even efficiently computable. In our work on the computability of
conditioning that came up earlier in the discussion, we show that adding sufficiently
smooth independent noise to a random variable allows us to perform conditioning in
situations where we would not have been able to otherwise. There are examples of this idea
elsewhere. For example, Braverman, Grigo, and Rojas <http://arxiv.org/abs/1201.0488> study
noise and intractability in dynamical systems. Specifically, they show that computing the
invariant measure characterizing the long-term statistical behavior of dynamical systems
is not possible. The road block is the computational power of the dynamical system itself.
The addition of a small amount of noise to the dynamics, however, decreases the
computational power of the dynamical system, and suffices to make the invariant measure
computable. In a world subject to noise (or, at least, well modeled as such), it seems
that many theoretical obstructions melt away.
------------------------------------------------------------------------------------------
*Luke*: Thanks, Daniel!
The post Daniel Roy on probabilistic programming and AI
<http://intelligence.org/2014/09/04/daniel-roy/> appeared first on Machine Intelligence
Research Institute <http://intelligence.org>.
You are subscribed to email updates from Machine Intelligence Research Institute » Blog
<http://intelligence.org>
To stop receiving these emails, you may unsubscribe now
<http://feedburner.google.com/fb/a/mailunsubscribe?k=HYpy2sNsssu2Xt-WF4TaHmHlEas>. Email
delivery powered by Google
Google Inc., 20 West Kinzie, Chicago IL USA 60610
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
