Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
text
Languages:
English
Size:
10K - 100K
License:
{"forum": "S1gc4XF8Lr", "submission_url": "https://openreview.net/forum?id=S1gc4XF8Lr", "submission_content": {"TL;DR": "We study the structure of noise in the brain and find it may help generalization by moving representations along in-class stimulus variations.", "keywords": ["noise", "trial-to-trial variability", "subspace", "generalization", "dropout"], "authors": ["Brian Hu", "Jiaqi Shang", "Ramakrishnan Iyer", "Josh Siegle", "Stefan Mihalas"], "title": "Does the neuronal noise in cortex help generalization?", "abstract": "Neural activity is highly variable in response to repeated stimuli. We used an open dataset, the Allen Brain Observatory, to quantify the distribution of responses to repeated natural movie presentations. A large fraction of responses are best fit by log-normal distributions or Gaussian mixtures with two components. These distributions are similar to those from units in deep neural networks with dropout. Using a separate set of electrophysiological recordings, we constructed a population coupling model as a control for state-dependent activity fluctuations and found that the model residuals also show non-Gaussian distributions. We then analyzed responses across trials from multiple sections of different movie clips and observed that the noise in cortex aligns better with in-clip versus out-of-clip stimulus variations. We argue that noise is useful for generalization when it moves along representations of different exemplars in-class, similar to the structure of cortical noise.", "authorids": ["brianh@alleninstitute.org", "jiaqi.shang@alleninstitute.org", "rami@alleninstitute.org", "joshs@alleninstitute.org", "stefanm@alleninstitute.org"], "pdf": "/pdf/530499ff8cf42f3c67535d1520d47a1680b24b5c.pdf", "paperhash": "hu|does_the_neuronal_noise_in_cortex_help_generalization"}, "submission_cdate": 1568211762122, "submission_tcdate": 1568211762122, "submission_tmdate": 1572587831479, "submission_ddate": null, "review_id": ["HklQnwfwPr", "rJlgG_TPwS", "BkghhGU5DB"], "review_url": ["https://openreview.net/forum?id=S1gc4XF8Lr¬eId=HklQnwfwPr", "https://openreview.net/forum?id=S1gc4XF8Lr¬eId=rJlgG_TPwS", "https://openreview.net/forum?id=S1gc4XF8Lr¬eId=BkghhGU5DB"], "review_cdate": [1569298347420, 1569343496454, 1569510067588], "review_tcdate": [1569298347420, 1569343496454, 1569510067588], "review_tmdate": [1570047560710, 1570047558215, 1570047547762], "review_readers": [["everyone"], ["everyone"], ["everyone"]], "review_writers": [["NeurIPS.cc/2019/Workshop/Neuro_AI/Paper52/AnonReviewer2"], ["NeurIPS.cc/2019/Workshop/Neuro_AI/Paper52/AnonReviewer3"], ["NeurIPS.cc/2019/Workshop/Neuro_AI/Paper52/AnonReviewer1"]], "review_reply_count": [{"replyCount": 0}, {"replyCount": 0}, {"replyCount": 0}], "review_replyto": ["S1gc4XF8Lr", "S1gc4XF8Lr", "S1gc4XF8Lr"], "review_content": [{"evaluation": "2: Poor", "intersection": "3: Medium", "importance_comment": "CNNs with dropout are proposed as a null model which seems to explain the response noise in neural data, but CNNs have no eye movements, while the neural data apparently does. ", "clarity": "2: Can get the general idea", "technical_rigor": "2: Marginally convincing", "intersection_comment": "While the analysis of the noise properties of neural responses in these datasets is important for the field of Neuroscience, descriptions of the datasets used and control for eye movements seems to be lacking. This work appears to offer little new insight for the AI field in its current preliminary form.", "rigor_comment": "deltaF/F - explain.\nFor neuroAI symposium, better to define these terms.\n\nDetails on convnet too sparse. Maxpooling layers? What was achieved performance on CIFAR-10 test set?\n\nI would liked to have seen a log-norm example class in Fig 1.\nThe mixture of Gaussians depicted in Fig 1 (a), is one to consider <0.1 signal, or noise? Perhaps its no response plus noise? Do you have any baseline \"no stimulus\" epochs to quantify the level of baseline noise for each of the units considered?\n\nFor Allen Institute data, and the head fixed mice, apparently eye movements were not paralyzed.\nThis is an important source of variability which is not accounted for in the CNN model.\ni.e. for a given presentation of the stimulus, one has no idea if receptive fields are receiving remotely similar illumination.\nPerhaps a way to proceed to control for this would be to inject the Allen Institute recorded eye movements into the CNN model input stream.\n\n", "comment": "AI->Neuro:\n\"We believe that research into the structure and role of biological noise will be useful for developing new methods to train neural networks with better generalization capabilities.\"\nThe inspiration appears to have propagated in the opposite direction, i.e. CNNs with dropout are proposed as a null model which seems to explain the observations in the data.\n\n\"Future work will study how different forms of subspace-aligned noise may help deep neural networks generalize better from fewer examples.\"\nThis is an interesting prospect of the work proposed in the abstract, but essentially left for future investigation.\n\nWith this in mind, while the analysis of the noise properties of neural reponses in these datasets is important for the field of Neuroscience, descriptions of the datasets used and control for eye movements seems to be lacking. This work appears to offer little new insight for the AI field.\n", "importance": "2: Marginally important", "title": "Dropout inspires a hypothesis on the functional role for noise in generalization, but analysis is superficial.", "category": "AI->Neuro", "clarity_comment": "See rigor."}, {"evaluation": "1: Very poor", "intersection": "2: Low", "importance_comment": "Tries to argue based on which distributions fit best the noise measured in cortex that the networks are similar to ANNs with dropout. However, the link is tenuous and not strongly argued.", "clarity": "2: Can get the general idea", "technical_rigor": "1: Not convincing", "intersection_comment": "In principle showing that dropout was like neuronal noise could be interesting, but it's pretty tenuous here.", "rigor_comment": "It's a weak link to show that two distributions are similar, especially when only a very few distributions were fit and there are a lot of missing details about what is meant by a \"dropout-like\" distribution (early on it says \"see Methods\" but then isn't defined in the Methods).", "comment": "There are a number of claims that don't appear to be very well supported by what is actually shown.", "importance": "1: Irrelevant", "title": "Tries to link observations of noise in cortex to dropout, but relationship is unclear", "category": "AI->Neuro", "clarity_comment": "Quite difficult to follow the chain of reasoning here."}, {"title": "Quantification of neural noise is poorly described, and alternative explanations not explored", "importance": "1: Irrelevant", "importance_comment": "The authors suggest that noise corresponds to a regularization step that the brain does, like dropout. Single neuron noise could result in information from that neuron not propagating further down the network, like in dropout (but not equivalent for various reasons), but that would require more thoughtful comparisons with the neural activity, instead of just a comparison of distribution of activations.", "rigor_comment": "Controlling for state variables was a useful step to perform. Otherwise the comparisons are not well-quantified and other explanations of the noise/models are not explored.", "clarity_comment": "Figure 3 needs much more description, are the subspaces defined using trial-averaged responses? Why define a noise subspace instead of looking at the direction of the noise on each trial?", "clarity": "3: Average readability", "evaluation": "2: Poor", "intersection_comment": "Comparisons are too preliminary.", "intersection": "2: Low", "technical_rigor": "2: Marginally convincing", "category": "AI->Neuro"}], "comment_id": [], "comment_cdate": [], "comment_tcdate": [], "comment_tmdate": [], "comment_readers": [], "comment_writers": [], "comment_reply_content": [], "comment_content": [], "comment_replyto": [], "comment_url": [], "meta_review_cdate": null, "meta_review_tcdate": null, "meta_review_tmdate": null, "meta_review_ddate ": null, "meta_review_title": null, "meta_review_metareview": null, "meta_review_confidence": null, "meta_review_readers": null, "meta_review_writers": null, "meta_review_reply_count": null, "meta_review_url": null, "decision": "Accept (Poster)"} |