We theoretically study the fundamental problem of learning a single neuron with a bias term ($\mathbf{x} \mapsto \sigma(<\mathbf{w},\mathbf{x}> + b)$) in the realizable setting with the ReLU activation, using gradient descent. Perhaps surprisingly, we show that this is a significantly different and more challenging problem than the bias-less case (which was the focus of previous works on single neurons), both in terms of the optimization geometry as well as the ability of gradient methods to succeed in some scenarios. We provide a detailed study of this problem, characterizing the critical points of the objective, demonstrating failure cases, and providing positive convergence guarantees under different sets of assumptions. To prove our results, we develop some tools which may be of independent interest, and improve previous results on learning single neurons.
Several recent works have shown separation results between deep neural networks, and hypothesis classes with inferior approximation capacity such as shallow networks or kernel classes. On the other hand, the fact that deep networks can efficiently express a target function does not mean this target function can be learned efficiently by deep neural networks. In this work we study the intricate connection between learnability and approximation capacity. We show that learnability with deep networks of a target function depends on the ability of simpler classes to approximate the target. Specifically, we show that a necessary condition for a function to be learnable by gradient descent on deep neural networks is to be able to approximate the function, at least in a weak sense, with shallow neural networks. We also show that a class of functions can be learned by an efficient statistical query algorithm if and only if it can be approximated in a weak sense by some kernel class. We give several examples of functions which demonstrate depth separation, and conclude that they cannot be efficiently learned, even by a hypothesis class that can efficiently approximate them.
Graph neural networks (GNNs) can process graphs of different sizes but their capacity to generalize across sizes is still not well understood. Size generalization is key to numerous GNN applications, from solving combinatorial optimization problems to learning in molecular biology. In such problems, obtaining labels and training on large graphs can be prohibitively expensive, but training on smaller graphs is possible. This paper puts forward the size-generalization question and characterizes important aspects of that problem theoretically and empirically. We show that even for very simple tasks, GNNs do not naturally generalize to graphs of larger size. Instead, their generalization performance is closely related to the distribution of patterns of connectivity and features and how that distribution changes from small to large graphs. Specifically, we show that in many cases, there are GNNs that can perfectly solve a task on small graphs but generalize poorly to large graphs and that these GNNs are encountered in practice. We then formalize size generalization as a domain-adaption problem and describe two learning setups where size generalization can be improved. First, as a self-supervised learning problem (SSL) over the target domain of large graphs. Second, as a semi-supervised learning problem when few samples are available in the target domain. We demonstrate the efficacy of these solutions on a diverse set of benchmark graph datasets.
We study the effects of mild over-parameterization on the optimization landscape of a simple ReLU neural network of the form $\mathbf{x}\mapsto\sum_{i=1}^k\max\{0,\mathbf{w}_i^{\top}\mathbf{x}\}$, in a well-studied teacher-student setting where the target values are generated by the same architecture, and when directly optimizing over the population squared loss with respect to Gaussian inputs. We prove that while the objective is strongly convex around the global minima when the teacher and student networks possess the same number of neurons, it is not even \emph{locally convex} after any amount of over-parameterization. Moreover, related desirable properties (e.g., one-point strong convexity and the Polyak-{\L}ojasiewicz condition) also do not hold even locally. On the other hand, we establish that the objective remains one-point strongly convex in \emph{most} directions (suitably defined). For the non-global minima, we prove that adding even just a single neuron will turn a non-global minimum into a saddle point. This holds under some technical conditions which we validate empirically. These results provide a possible explanation for why recovering a global minimum becomes significantly easier when we over-parameterize, even if the amount of over-parameterization is very moderate.
We consider the fundamental problem of learning a single neuron $x \mapsto\sigma(w^\top x)$ using standard gradient methods. As opposed to previous works, which considered specific (and not always realistic) input distributions and activation functions $\sigma(\cdot)$, we ask whether a more general result is attainable, under milder assumptions. On the one hand, we show that some assumptions on the distribution and the activation function are necessary. On the other hand, we prove positive guarantees under mild assumptions, which go beyond those studied in the literature so far. We also point out and study the challenges in further strengthening and generalizing our results.
The lottery ticket hypothesis (Frankle and Carbin, 2018), states that a randomly-initialized network contains a small subnetwork such that, when trained in isolation, can compete with the performance of the original network. We prove an even stronger hypothesis (as was also conjectured in Ramanujan et al., 2019), showing that for every bounded distribution and every target network with bounded weights, a sufficiently over-parameterized neural network with random weights contains a subnetwork with roughly the same accuracy as the target network, without any further training.
Recently, a spate of papers have provided positive theoretical results for training over-parameterized neural networks (where the network size is larger than what is needed to achieve low error). The key insight is that with sufficient over-parameterization, gradient-based methods will implicitly leave some components of the network relatively unchanged, so the optimization dynamics will behave as if those components are essentially fixed at their initial random values. In fact, fixing these explicitly leads to the well-known approach of learning with random features. In other words, these techniques imply that we can successfully learn with neural networks, whenever we can successfully learn with random features. In this paper, we first review these techniques, providing a simple and self-contained analysis for one-hidden-layer networks. We then argue that despite the impressive positive results, random feature approaches are also inherently limited in what they can explain. In particular, we rigorously show that random features cannot be used to learn even a single ReLU neuron with standard Gaussian inputs, unless the network size (or magnitude of the weights) is exponentially large. Since a single neuron is learnable with gradient-based methods, we conclude that we are still far from a satisfying general explanation for the empirical success of neural networks.