oergiR

One of the complaints from the mathematician is that OpenAI cannot tell whether his data has been used as training data. Not many people realise this is a direct consequence of the GDPR.

The GDPR protects PII, personally identifiable information, and the definition of PII does not include “mathematics that only this person can think of”. As long as OpenAI strips out PII and removes identifiers linking the conversation to a person, the GDPR is happy. Without the GDPR, OpenAI might have kept the identifiers with the data, and been able to say whether a specific conversation was in the training data.

There is more to this quote than you might think.

Grammatically, in English the verb "swim" requires an "animate subject", i.e. a living being, like a human or an animal. So the question of whether a submarine can swim is about grammar. In Russian (IIRC), submarines can swim just fine, because the verb does not have this animacy requirement. Crucially, the question is not about whether or how a submarine propels itself.

Likewise, in English at least, the verb "think" requires an animate object. the question whether a machine can think is about whether you consider it to be alive. Again, whether or how the machine generates its output is not material to the question.

FreeType was written when fonts were local, trusted, resources, and it was written in low-level C to be fast. The TrueType/OpenType format is also made for fast access, e.g. with internal pointers, making validation a pain.

So though FreeType is carefully written w.r.t. correctness, it was not meant to deal with malicious input and that robustness is hard to put in afterwards.

That makes it sound like it’s a choice, which it isn’t really. The way to look at it is from a probabilistic perspective: with the fix, you maximise the probability of the data. Without the fix, you fairly arbitrarily raise some probabilities to a power greater than one, and some to a power less than one.

Isn’t the main problem the cost of heating? In the UK at least a kWh of electricity cost 4 times what a kWh of gas costs. Even if a heat pump is twice as efficient as gas heating, the cost of heating is still twice as high.

I agree except for (6). A language model assigns probabilities to sequences. The model needs normalised distributions, eg using a softmax, so that’s the right way of thinking about it.

In English everywhere except the US, it’s “horse riding”; “horseback riding” is US English.

Definitely humorous: https://m.youtube.com/watch?v=5wSw3IWRJa0

Training a speaker-specific recogniser that improves over a generic recogniser requires a lot more data nowadays. First, generic systems are a lot better and trained on a lot more data nowadays. Second, speaker adaptation worked better for the Gaussian mixture models from the late nineties (don’t know about the eighties) than for neural networks.

The actual paper is available here: http://www.isca-speech.org/archive/Interspeech_2017/abstract...

This is how healthcare is set up in Switzerland and the Netherlands.

It is just an implementation issue how payments (through tax or otherwise) are routed. In a system that involves market forces, the key thing is that if everybody is insured, including the healthy, the cost per person comes down.

The "model" in the title is the model of the world, as a probabilistic model. The good thing about such a model is that it explicitly states your beliefs about the world. Once you've defined it, in theory reasoning about it is straightforward. (In practice a lot of papers get written about how to do approximate inference.) It's also straightforward to do unsupervised learning.

This is a different perspective from (most uses of) neural networks, which do not have this clear separation between the model and how to reason about it. It's funny that Chris Bishop in 1995 wrote the textbook "Neural Networks for Pattern Recognition" and now is effectively arguing against using neural networks.

You can use both by using neural networks as "factors" (the black squares) in probabilistic models.

To train a speech recogniser on subtitles, you need to be careful to allow for words missing from the transcriptions. But it has been done: goo.gl/Syx27j

I know that Andrew Ng and colleagues say that they don't use HMMs. I haven't spoken with them (I haven't seen them at speech conferences) so I do not know whether they actually believe this themselves.

I believe the best comparison between "CTC" (which is billed as recurrent neural networks without the HMMs) and the traditional approach is by people at Google, Sak et al, "Learning Acoustic frame labeling for speech recognition with recurrent neural networks", ICASSP 2015. (I can't find a PDF online.)

Probabilistic models. Recent research often focuses on Bayesian models.

Probabilistic models have never really gone away. This presentation by LeCun actually suggests embedding neural networks inside of various types of probabilistic models: factor graphs and conditional random fields. This is, for example, how speech recognition works: the output of a neural network is fed into a probabilistic model (a hidden Markov model).