Derivatives play a central function in optimization and machine studying. By domestically approximating a coaching loss, derivatives information an optimizer towards decrease values of the loss. Automatic differentiation frameworks comparable to TensorFlow, PyTorch, and JAX are a necessary a part of fashionable machine studying, making it possible to make use of gradient-based optimizers to coach very advanced fashions.
But are derivatives all we want? By themselves, derivatives solely inform us how a perform behaves on an infinitesimal scale. To use derivatives successfully, we regularly have to know greater than that. For instance, to decide on a studying charge for gradient descent, we have to know one thing about how the loss perform behaves over a small however finite window. A finite-scale analogue of automatic differentiation, if it existed, might assist us make such selections extra successfully and thereby velocity up coaching.
In our new paper “Automatically Bounding The Taylor Remainder Series: Tighter Bounds and New Applications”, we current an algorithm referred to as AutoBound that computes polynomial higher and decrease bounds on a given perform, that are legitimate over a user-specified interval. We then start to discover AutoBound’s functions. Notably, we current a meta-optimizer referred to as SafeRate that makes use of the higher bounds computed by AutoBound to derive studying charges which are assured to monotonically scale back a given loss perform, with out the necessity for time-consuming hyperparameter tuning. We are additionally making AutoBound out there as an open-source library.
The AutoBound algorithm
Given a perform f and a reference level x0, AutoBound computes polynomial higher and decrease bounds on f that maintain over a user-specified interval referred to as a belief area. Like Taylor polynomials, the bounding polynomials are equal to f at x0. The bounds grow to be tighter because the belief area shrinks, and strategy the corresponding Taylor polynomial because the belief area width approaches zero.
 |
| Automatically-derived quadratic higher and decrease bounds on a one-dimensional perform f, centered at x0=0.5. The higher and decrease bounds are legitimate over a user-specified belief area, and grow to be tighter because the belief area shrinks. |
Like automatic differentiation, AutoBound will be utilized to any perform that may be applied utilizing commonplace mathematical operations. In reality, AutoBound is a generalization of Taylor mode automatic differentiation, and is equal to it within the particular case the place the belief area has a width of zero.
To derive the AutoBound algorithm, there have been two primary challenges we needed to tackle:
- We needed to derive polynomial higher and decrease bounds for numerous elementary features, given an arbitrary reference level and arbitrary belief area.
- We needed to provide you with an analogue of the chain rule for combining these bounds.
Bounds for elementary features
For quite a lot of commonly-used features, we derive optimum polynomial higher and decrease bounds in closed type. In this context, “optimum” means the bounds are as tight as potential, amongst all polynomials the place solely the maximum-degree coefficient differs from the Taylor collection. Our concept applies to elementary features, comparable to exp and log, and customary neural community activation features, comparable to ReLU and Swish. It builds upon and generalizes earlier work that utilized solely to quadratic bounds, and just for an unbounded belief area.
 |
| Optimal quadratic higher and decrease bounds on the exponential perform, centered at x0=0.5 and legitimate over the interval [0, 2]. |
A brand new chain rule
To compute higher and decrease bounds for arbitrary features, we derived a generalization of the chain rule that operates on polynomial bounds. To illustrate the concept, suppose we have now a perform that may be written as
and suppose we have already got polynomial higher and decrease bounds on g and h. How will we compute bounds on f?
The key seems to be representing the higher and decrease bounds for a given perform as a single polynomial whose highest-degree coefficient is an interval relatively than a scalar. We can then plug the sure for h into the sure for g, and convert the end result again to a polynomial of the identical type utilizing interval arithmetic. Under appropriate assumptions in regards to the belief area over which the sure on g holds, it may be proven that this process yields the specified sure on f.
 |
| The interval polynomial chain rule utilized to the features h(x) = sqrt(x) and g(y) = exp(y), with x0=0.25 and belief area [0, 0.5]. |
Our chain rule applies to one-dimensional features, but in addition to multivariate features, comparable to matrix multiplications and convolutions.
Propagating bounds
Using our new chain rule, AutoBound propagates interval polynomial bounds by means of a computation graph from the inputs to the outputs, analogous to forward-mode automatic differentiation.
 |
| Forward propagation of interval polynomial bounds for the perform f(x) = exp(sqrt(x)). We first compute (trivial) bounds on x, then use the chain rule to compute bounds on sqrt(x) and exp(sqrt(x)). |
To compute bounds on a perform f(x), AutoBound requires reminiscence proportional to the dimension of x. For this cause, sensible functions apply AutoBound to features with a small variety of inputs. However, as we’ll see, this doesn’t stop us from utilizing AutoBound for neural community optimization.
Automatically deriving optimizers, and different functions
What can we do with AutoBound that we could not do with automatic differentiation alone?
Among different issues, AutoBound can be utilized to mechanically derive problem-specific, hyperparameter-free optimizers that converge from any start line. These optimizers iteratively scale back a loss by first utilizing AutoBound to compute an higher sure on the loss that’s tight on the present level, after which minimizing the higher sure to acquire the following level.
 |
| Minimizing a one-dimensional logistic regression loss utilizing quadratic higher bounds derived mechanically by AutoBound. |
Optimizers that use higher bounds on this manner are referred to as majorization-minimization (MM) optimizers. Applied to one-dimensional logistic regression, AutoBound rederives an MM optimizer first printed in 2009. Applied to extra advanced issues, AutoBound derives novel MM optimizers that will be tough to derive by hand.
We can use an identical concept to take an current optimizer comparable to Adam and convert it to a hyperparameter-free optimizer that’s assured to monotonically scale back the loss (within the full-batch setting). The ensuing optimizer makes use of the identical replace route as the unique optimizer, however modifies the training charge by minimizing a one-dimensional quadratic higher sure derived by AutoBound. We confer with the ensuing meta-optimizer as SafeRate.
 |
| Performance of SafeRate when used to coach a single-hidden-layer neural community on a subset of the MNIST dataset, within the full-batch setting. |
Using SafeRate, we are able to create extra sturdy variants of current optimizers, at the price of a single extra ahead move that will increase the wall time for every step by a small issue (about 2x within the instance above).
In addition to the functions simply mentioned, AutoBound can be utilized for verified numerical integration and to mechanically show sharper variations of Jensen’s inequality, a elementary mathematical inequality used regularly in statistics and different fields.
Improvement over classical bounds
Bounding the Taylor the rest time period mechanically will not be a brand new concept. A classical approach produces diploma okay polynomial bounds on a perform f which are legitimate over a belief area [a, b] by first computing an expression for the okayth spinoff of f (utilizing automatic differentiation), then evaluating this expression over [a,b] utilizing interval arithmetic.
While elegant, this strategy has some inherent limitations that may result in very unfastened bounds, as illustrated by the dotted blue traces within the determine beneath.
 |
| Quadratic higher and decrease bounds on the lack of a multi-layer perceptron with two hidden layers, as a perform of the preliminary studying charge. The bounds derived by AutoBound are a lot tighter than these obtained utilizing interval arithmetic analysis of the second spinoff. |
Looking ahead
Taylor polynomials have been in use for over 300 years, and are omnipresent in numerical optimization and scientific computing. Nevertheless, Taylor polynomials have important limitations, which may restrict the capabilities of algorithms constructed on high of them. Our work is a part of a rising literature that acknowledges these limitations and seeks to develop a brand new basis upon which extra sturdy algorithms will be constructed.
Our experiments up to now have solely scratched the floor of what’s potential utilizing AutoBound, and we imagine it has many functions we have now not found. To encourage the analysis neighborhood to discover such potentialities, we have now made AutoBound out there as an open-source library constructed on high of JAX. To get began, go to our GitHub repo.
Acknowledgements
This put up relies on joint work with Josh Dillon. We thank Alex Alemi and Sergey Ioffe for helpful suggestions on an earlier draft of the put up.