7 ms·
Cynthia Rudin and interpretable ML models
- derbOac 3y agoI've always seen interpretability and explainability as different sides of the same coin. If you take an information-theoretic approach to it, and think of a DL model like any other model, there is a certain equivalence in understanding the model features and how it behaves with reference to the universe of data it is applied to. It was an interesting article but I felt like it created problems that need not be there (or maybe it's just describing problems that others created?)
- ImprobableTruth 3y agoNot sure if it was intentional,but the title of this article is pretty hilarious considering she explicitly badmouths explainability (trying to peer inside a black box) and advocates for interpretability (building models that are less black-boxy in nature). Also, man, quanta feels really rough and popsci-y when it comes to CS.
- OnlineGladiator 3y ago> Also, man, quanta feels really rough and popsci-y when it comes to CS. I'm guessing this is just the Gell-Mann Amnesia effect. Why do you think the quality is better for other fields?
- gwern 3y agoI was amused by her example of poetry needing interpretability, given that we already know why and did when they did that research: https://gwern.net/gpt-3#bpes https://gwern.net/gpt-3#bpes
- janalsncm 3y agoI’ve always wondered what an sufficient explanation of a neural network would entail. At a very low level, there’s no secret: all of the weights are there, all of the relationships between weights are known. The problem is that it doesn’t tell you anything about the emergent properties of the network, in the same way that quantum physics doesn’t give much insight into biology. It may be possible that there is no English sentence you can utter about the network which is both explanatory and fully accurate. What is the network doing? It’s trying to approximate the function you’ve given it. That’s it. You can try other things like ablation to find the effects of lobotomizing the network in certain ways, but this also can’t fully explain 2nd and higher order relationships.
- sdwr 3y agoSame thing as the brain pretty much, except we have root control over the NN in a way we don't with the brain. To my knowledge, some angles have been explored. Finding which inputs maximally stimulate indiv neurons, and tracing that up the chain. There was a fantastic article that did this on ImageNet - the takeaway was that neurons adjacent to the input encode sharp, basic features (like edges), and "higher-level" ones encode more nuanced stuff (textures, fur, wheels). Then ablations, as you mentioned. And finally, rerunning the same training data on differently sized + shaped architectures. AI systems are missing at least one level of complexity. All neurons in a NN layer fire simultaneously, unlike brain neurons, which trigger each other async.
- janalsncm 3y agoThe brain is heterogenous in ways that a neural network needn’t be. Certain brain regions definitely perform certain functions, certain regions are hardwired to other regions. You can chop out a chunk of a person’s brain and they’ll be completely unable to do certain things. Afaik the properties you’re referring to on AlexNet (the network trained to perform on Image Net) have to do with the nature of repeated convolution operations which, while interesting, is not as deeply insightful as I believe the article is aiming for.
- obviouslynotme 3y ago>It may be possible that there is no English sentence you can utter about anything which is both explanatory and fully accurate. FTFY. Philosophy aside, humans understand complex things by building simpler abstractions for them. I think the key to NN understandability centers on building simpler NNs focused on specific tasks that build into bigger ones.
- kypro 3y agoThis sentence from the article stood out to me: > The explanations have to be wrong, because if their explanations were always right, you could just replace the black box with the explanations. I've never thought of it like that, but I think that really gets to the core of the issue... To be able to reason about the behaviour of a neural network with 100% confidence you would need an explanation that incorporated every weight, otherwise you have to accept some degree of uncertainty. But the idea you could ever explain something with billions of parameters in a way that a human could comprehend seems ridiculous. Imperfect generalisation therefore would inevitably be required – "this bit of the network seems to perform x function". But in doing this you have to accept that this generalisation comes with uncertainty because such a high-level approximation is unlikely to capture all of the nuances. And if it did capture all nuances then the black box isn't needed.
- triyambakam 3y ago> They extract deeply hidden patterns in large data sets that our limited human brains can’t parse. I think that the bulk of ML has so far produced what our brains in fact easily see. We can easily perform classification, or generation.
- dboreham 3y agoThey're definitely worse than me at spotting spam.
- HyperSane 3y agoWe have no idea how the human brain works but no one seems to care.
- ChatGTP 3y agoThe only ingesting things now are the machines, that’s where all the money goes now. We’re just “meat bags”.
- dang 3y agoCan you please stop posting unsubstantive comments? This is below the quality line for an interesting HN thread.
- edgefield 3y agoI’m pretty sure the field of neuroscience is focused on understanding how the brain works.
- HyperSane 3y agoHave we made any fundamental progress in understanding how the human brain works in the last 20 years? 50?
- famouswaffles 3y agoWe've made pretty good progress. But the same problem that plagues explainabilty of ANN's rears its ugly head in neuroscience. Low level descriptions provide shockingly little insight into high level behavior.
- Blammar 3y agoI view explainability or interpretability of a network as the ability to take a network and replace it with a drastically smaller set of functions and tables that (a) you can explain and (b) work pretty much the same as the network does. Because we understand these functions and tables, we understand exactly how well the network will work, and also what is missing (i.e., how we can expand its accuracy.) I think this is a very hard problem, but it is one that needs to be solved.
- RosanaAnaDana 3y agoSometimes I think the interpretation of model parameters is all bunk. I think the legacy of over interpretation of parameters and results has resulted in the occification of some very shaky science.
- armchairhacker 3y agoGetting AI to show its work isn't just for accountability. "Showing your work" gives you a clearer picture of the problem / solution and prevents / fixes bugs in implicit reasoning, the key problem current AI has which prevents it from being truly autonomous. Ask GPT4 to do a task, and then ask it to the same task showing its work; you'll find that GPT4 is less likely to make mistakes on the latter. This is especially apparent for tasks like counting # of words and multi-step problems, which GPT normally has trouble with. But GPT4 still tends to struggle even breaking the task down, to the point where it starts producing extremely obvious mistakes (e.g. "the turtle moves 1 unit up, from (1, 0) to (2, 0)"). One possibility is that it isn't actually showing its work, it's just generating backwards explanations from a latent conclusion. Maybe this research will clarify whether this is the case, and help us develop a more coherent LLM.
- barking_biscuit 3y ago>If you want to trust a prediction, you need to understand how all the computations work. I disagree with the premise here. You don't understand how all the computations work in the brains of people's predictions whom you trust. You simply have a mental calculation of their batting average through exposure to their track record and this batting average functions as a proxy for trust. I find this is more or less the same way that I learn whether or not I can rely on GPT-4 for a particular use-case. If it's batting average is north of a certain % for a given use-case, then it doesn't need to be right 100% of the time for me to derive value from relying on it. I think we are slowly crossing a threshold where we accept indeterminism and mistakes from machines in a way that we haven't in the past.
- uoaei 3y agoThey may be referring more to a theory of mind than a theory of neuro-physics. I trust the decisions and opinions of those whose reasoning capacities I trust. It's not enough to know what they decide, you must also know the set of choices in front of them at the time and their process for choosing one among them.
- inimino 3y agoThe point is that you trust those people. You can't trust a black-box model, but you can trust a result if the explanation that is provided actually explains it. That is, if the explanation has factors X, Y, and Z, then every time those hold the result should be the same. Otherwise, you are just trusting the training set and the training process. Sometimes that's fine, as the article also mentions.
- barking_biscuit 3y agoNo, the point is those people are also a black-box to you and they are a black-box that you trust because of YOUR training set and training process.
- inimino 3y agoYes, your exact thought process may be a black box to me, but I myself am a black box made to the same (DNA etc) spec. We have millions of ancestral years of experience trusting each other and we know how to simulate each other because we are the same kind of black box. Your trust in other humans is explained much more by your priors over human behavior than by any observed behavior of individual humans--that's layered on top. In short: we trust people because we can model them effectively, and being one helps. We trust mathematical models and traditional programs because we understand them and can check their work in any specific instance. Large ANNs don't (yet) benefit from either of these two kinds of trust.
- graycat 3y agoThe models, neural networks Professor Rudin is considering have a LOT of parameters, dimensions, neurons, neuron values, etc. Okay. Since apparently no one has the explanations desired, we have to guess. So, let's do some guessing: Given so many parameters, etc., we have in some sense -- in some case of geometry, spaces, maybe vector spaces, maybe as in linear algebra -- a lot of dimensions. Then something surprising holds (once we get precise about a space, easy enough to prove): Given a sphere in the space, we can calculate its volume. We can do this for the space of any finite dimension. Here is the surprise: As we have a lot of dimensions, there is a LOT of volume in that sphere, and nearly all that volume is just inside the surface of that sphere. E.g., if do some work in nearest neighbors, discover this surprise in strong terms. Net, in the space being considered, there is a LOT of volume. Then ...: There is plenty of volume to put faces of cats over here, dogs over there, men another place, women still well separated, essays on bone cancer far away, ..., for thousands, millions, ..., more things, thoughts, topics, etc. Then given some new data, say, a white cat not in the training data, likely the data on that white cat will settle on the volume with the cats instead of dogs, monkeys, etc. and, thus, we will have recognized a cat via some emergent functionality. Just a guess.
- PeterStuer 3y agoTo trust a prediction you do not need to understand the underlying computations. What you do need is an on demand understandable rational justifucation of how the prediction was derived at the right semantic domain level.