16 results on '"Roberts, Lindon"'
Search Results
2. On the selection of the weighting parameter value in optimizing Eucalyptus globulus pulp yield models based on NIR spectra
- Author
-
Zhen, Yi, Ho, Tu X., Roberts, Lindon, Schimleck, Laurence R., and Sinha, Arijit
- Published
- 2022
- Full Text
- View/download PDF
3. Derivative-free algorithms for nonlinear optimisation problems
- Author
-
Roberts, Lindon, Cartis, Coralia, Fiala, Jan, and Marteau, Benjamin
- Subjects
518 ,Mathematical optimization ,Numerical analysis - Abstract
Minimising a nonlinear function is a ubiquitous problem in applications, and standard algorithms need accurate evaluations of the function and its derivatives. However, if the function is black-box, expensive to evaluate, or noisy, it may be impractical or even impossible to obtain accurate derivatives, and we require derivative-free optimisation (DFO). Model-based DFO methods perform well in practice, taking many features from derivative-based methods, but can have lower performance for noisy problems and a high linear algebra cost. In this thesis, we aim to improve the flexibility, robustness and scalability of state-of-the-art methods for model-based DFO by designing, analysing, implementing and comprehensively testing three new algorithms for nonlinear optimisation, with a special focus on nonlinear least-squares problems (such as parameter fitting). The first, DFO-LS, is designed for nonlinear least-squares problems, and is simpler than existing methods, constructing linear residual models with a flexible approach. DFO-LS can benefit from a reduced initialisation cost, and has improved scalability and runtime over existing methods without a performance penalty. DFO-LS also has features for noisy problems, including a novel multiple restarts mechanism, which are low cost in evaluations and linear algebra, giving DFO-LS better performance compared to existing techniques for handling noise. We then introduce Py-BOBYQA, an extension of BOBYQA (Powell, 2009) with a similar multiple restarts mechanism to DFO-LS, amongst other new enhancements. It matches or outperforms BOBYQA and other state-of-the-art solvers on both smooth and noisy problems. We also propose a simple extension of Py-BOBYQA that may enable it to escape local minima and progress towards the global optima. Lastly, we introduce, for large-scale nonlinear least-squares problems, DFBGN, the first model-based DFO method to optimise in low-dimensional subspaces at each iteration. DFBGN has lower linear algebra costs, improved runtime, and hence improved performance on large-scale problems than DFO-LS, as it can perform many more iterations within a reasonable runtime limit. It can also make progress under very small evaluation budgets, with a low runtime cost.
- Published
- 2019
4. A simplified convergence theory for Byzantine resilient stochastic gradient descent
- Author
-
Roberts, Lindon and Smyth, Edward
- Published
- 2022
- Full Text
- View/download PDF
5. Inexact Derivative-Free Optimization for Bilevel Learning
- Author
-
Ehrhardt, Matthias J. and Roberts, Lindon
- Published
- 2021
- Full Text
- View/download PDF
6. A derivative-free Gauss–Newton method
- Author
-
Cartis, Coralia and Roberts, Lindon
- Published
- 2019
- Full Text
- View/download PDF
7. Analyzing inexact hypergradients for bilevel learning.
- Author
-
Ehrhardt, Matthias J and Roberts, Lindon
- Subjects
- *
AUTOMATIC differentiation , *BILEVEL programming , *MACHINE learning - Abstract
Estimating hyperparameters has been a long-standing problem in machine learning. We consider the case where the task at hand is modeled as the solution to an optimization problem. Here the exact gradient with respect to the hyperparameters cannot be feasibly computed and approximate strategies are required. We introduce a unified framework for computing hypergradients that generalizes existing methods based on the implicit function theorem and automatic differentiation/backpropagation, showing that these two seemingly disparate approaches are actually tightly connected. Our framework is extremely flexible, allowing its subproblems to be solved with any suitable method, to any degree of accuracy. We derive a priori and computable a posteriori error bounds for all our methods and numerically show that our a posteriori bounds are usually more accurate. Our numerical results also show that, surprisingly, for efficient bilevel optimization, the choice of hypergradient algorithm is at least as important as the choice of lower-level solver. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF
8. Analyzing Inexact Hypergradients for Bilevel Learning
- Author
-
Ehrhardt, Matthias J. and Roberts, Lindon
- Subjects
FOS: Computer and information sciences ,Computer Science - Machine Learning ,Optimization and Control (math.OC) ,FOS: Mathematics ,Mathematics - Optimization and Control ,Machine Learning (cs.LG) - Abstract
Estimating hyperparameters has been a long-standing problem in machine learning. We consider the case where the task at hand is modeled as the solution to an optimization problem. Here the exact gradient with respect to the hyperparameters cannot be feasibly computed and approximate strategies are required. We introduce a unified framework for computing hypergradients that generalizes existing methods based on the implicit function theorem and automatic differentiation/backpropagation, showing that these two seemingly disparate approaches are actually tightly connected. Our framework is extremely flexible, allowing its subproblems to be solved with any suitable method, to any degree of accuracy. We derive a priori and computable a posteriori error bounds for all our methods, and numerically show that our a posteriori bounds are usually more accurate. Our numerical results also show that, surprisingly, for efficient bilevel optimization, the choice of hypergradient algorithm is at least as important as the choice of lower-level solver.
- Published
- 2023
9. Optimizing illumination patterns for classical ghost imaging
- Author
-
Kingston, Andrew M., Roberts, Lindon, Aminzadeh, Alaleh, Pelliccia, Daniele, Svalbe, Imants D., and Paganin, David M.
- Subjects
Image and Video Processing (eess.IV) ,FOS: Electrical engineering, electronic engineering, information engineering ,FOS: Physical sciences ,Electrical Engineering and Systems Science - Image and Video Processing ,Optics (physics.optics) ,Physics - Optics - Abstract
Classical ghost imaging is a new paradigm in imaging where the image of an object is not measured directly with a pixelated detector. Rather, the object is subject to a set of illumination patterns and the total interaction of the object, e.g., reflected or transmitted photons or particles, is measured for each pattern with a single-pixel or bucket detector. An image of the object is then computed through the correlation of each pattern and the corresponding bucket value. Assuming no prior knowledge of the object, the set of patterns used to compute the ghost image dictates the image quality. In the visible-light regime, programmable spatial light modulators can generate the illumination patterns. In many other regimes -- such as x rays, electrons, and neutrons -- no such dynamically configurable modulators exist, and patterns are commonly produced by employing a transversely-translated mask. In this paper we explore some of the properties of masks or speckle that should be considered to maximize ghost-image quality, given a certain experimental classical ghost-imaging setup employing a transversely-displaced but otherwise non-configurable mask., 28 pages, 17 figures
- Published
- 2022
10. MODEL-BASED DERIVATIVE-FREE METHODS FOR CONVEX-CONSTRAINED OPTIMIZATION.
- Author
-
HOUGH, MATTHEW and ROBERTS, LINDON
- Subjects
- *
SET theory , *NONLINEAR equations , *INTERPOLATION - Abstract
We present a model-based derivative-free method for optimization subject to general convex constraints, which we assume are unrelaxable and accessed only through a projection operator that is cheap to evaluate. We prove global convergence and a worst-case complexity of O(ε-2) iterations and objective evaluations for nonconvex functions, matching results for the unconstrained case. We introduce new, weaker requirements on model accuracy compared to existing methods. As a result, sufficiently accurate interpolation models can be constructed only using feasible points. We develop a comprehensive theory of interpolation set management in this regime for linear and composite linear models. We implement our approach for nonlinear least-squares problems and demonstrate strong practical performance compared to general-purpose solvers. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
11. Escaping local minima with local derivative-free methods: a numerical investigation.
- Author
-
Cartis, Coralia, Roberts, Lindon, and Sheridan-Methven, Oliver
- Subjects
- *
GLOBAL optimization - Abstract
We investigate the potential of applying a state-of-the-art, local derivative-free solver, Py-BOBYQA to global optimization problems. In particular, we demonstrate the potential of a restarts procedure – as distinct from multistart methods – to allow Py-BOBYQA to escape local minima (where ordinarily it would terminate at the first local minimum found). We also introduce an adaptive variant of restarts which yields improved performance on global optimization problems. As Py-BOBYQA is a model-based trust-region method, we compare largely with other global optimization methods for which (global) models are important, such as Bayesian optimization and response surface methods; we also consider state-of-the-art representative deterministic and stochastic codes, such as DIRECT and CMA-ES. We find numerically that the restarts procedures in Py-BOBYQA are effective at helping it to escape local minima, when compared to using no restarts in Py-BOBYQA. Additionally, we find that Py-BOBYQA with adaptive restarts has comparable performance with global optimization solvers for all accuracy/budget regimes, in both smooth and noisy settings. In particular, Py-BOBYQA variants are best performing for smooth and multiplicative noise problems in high-accuracy regimes. As a by-product, some preliminary conclusions can be drawn on the relative performance of the global solvers we have tested with default settings. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
12. Scalable Derivative-Free Optimization for Nonlinear Least-Squares Problems
- Author
-
Cartis, Coralia, Ferguson, Tyler, and Roberts, Lindon
- Subjects
FOS: Computer and information sciences ,Computer Science - Machine Learning ,Optimization and Control (math.OC) ,FOS: Mathematics ,Mathematics - Optimization and Control ,Machine Learning (cs.LG) - Abstract
Derivative-free - or zeroth-order - optimization (DFO) has gained recent attention for its ability to solve problems in a variety of application areas, including machine learning, particularly involving objectives which are stochastic and/or expensive to compute. In this work, we develop a novel model-based DFO method for solving nonlinear least-squares problems. We improve on state-of-the-art DFO by performing dimensionality reduction in the observational space using sketching methods, avoiding the construction of a full local model. Our approach has a per-iteration computational cost which is linear in problem dimension in a big data regime, and numerical evidence demonstrates that, compared to existing software, it has dramatically improved runtime performance on overdetermined least-squares problems., Fixed author spelling
- Published
- 2020
13. Does Model Calibration Reduce Uncertainty in Climate Projections?
- Author
-
Tett, Simon F. B., Gregory, Jonathan M., Freychet, Nicolas, Cartis, Coralia, Mineter, Michael J., and Roberts, Lindon
- Subjects
CLIMATE sensitivity ,ATMOSPHERIC models ,CALIBRATION ,SURFACE temperature ,GENERAL circulation model - Abstract
Uncertainty in climate projections is large as shown by the likely uncertainty ranges in equilibrium climate sensitivity (ECS) of 2.5–4 K and in the transient climate response (TCR) of 1.4–2.2 K. Uncertainty in model projections could arise from the way in which unresolved processes are represented, the parameter values used, or the targets for model calibration. We show that, in two climate model ensembles that were objectively calibrated to minimize differences from observed large-scale atmospheric climatology, uncertainties in ECS and TCR are about 2–6 times smaller than in the CMIP5 or CMIP6 multimodel ensemble. We also find that projected uncertainties in surface temperature, precipitation, and annual extremes are relatively small. Residual uncertainty largely arises from unconstrained sea ice feedbacks. The more than 20-year-old HadAM3 standard model configuration simulates observed hemispheric-scale observations and preindustrial surface temperatures about as well as the median CMIP5 and CMIP6 ensembles while the optimized configurations simulate these better than almost all the CMIP5 and CMIP6 models. Hemispheric-scale observations and preindustrial temperatures are not systematically better simulated in CMIP6 than in CMIP5 although the CMIP6 ensemble seems to better simulate patterns of large-scale observations than the CMIP5 ensemble and the optimized HadAM3 configurations. Our results suggest that most CMIP models could be improved in their simulation of large-scale observations by systematic calibration. However, the uncertainty in climate projections (for a given scenario) likely largely arises from the choice of parameterization schemes for unresolved processes ("structural uncertainty"), with different tuning targets another possible contributor. Significance Statement: Climate models represent unresolved phenomena controlled by uncertain parameters. Changes in these parameters impact how well a climate model simulates current climate and its climate projections. Multiple calibrations of a single climate model, using an objective method, to large-scale atmospheric observations are performed. These models produce very similar climate projections at both global and regional scales. An analysis that combines uncertainties in observations with simulated sensitivity to observations and climate response also has small uncertainty showing that, for this model, current observations constrain climate projections. Recently developed climate models have a broad range of abilities to simulate large-scale climate with only some improvement in their ability to simulate this despite a decade of model development. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
14. A heat and mass transfer study of carbon paste baking.
- Author
-
Roberts, Lindon, Nordgård-Hansen, Ellen, Mikkelsen, Øyvind, Halvorsen, Svenn Anton, and Van Gorder, Robert A.
- Subjects
- *
MASS transfer , *HEAT transfer , *CARBON , *SMELTING furnaces , *ALUMINUM - Abstract
Ramming paste is a particular type of carbon paste which is used as lining for electric smelting furnaces and Hall-Héroult cells for the production of aluminium. The purpose of this lining is to form an impenetrable barrier, keeping the liquid within the furnace. If the lining has cracks or holes, then liquid can escape, which can lead to safety risks and financial losses, so the integrity of the lining is of great importance. In the present study, we develop a first principles mathematical model for the heat and mass transfer processes occurring during the baking of carbon paste. We then obtain numerical simulations using this model and compare the simulation results to experimental data, demonstrating that the model solutions do indeed describe and predict realistic behaviours of the carbon paste baking process. The simulations indicate a strong pressure buildup during the evaporation of water from fresh paste during the baking process, which is likely to lead to cracking of the paste as it hardens. Furthermore, we are able to show that more gradual heating during the baking process can lower the maximal pressures predicted by the model, which in turn may reduce the prevalence of cracks within the hardened paste. [ABSTRACT FROM AUTHOR]
- Published
- 2017
- Full Text
- View/download PDF
15. WHAT I GOT.
- Author
-
NOWELL, BRAD, Wilson, Eric, GAUGH, FLOYD, and ROBERTS, LINDON
- Abstract
The sheet music for the song "What I Got," by Sublime is presented.
- Published
- 2011
16. Mask design, fabrication, and experimental ghost imaging applications for patterned X-ray illumination.
- Author
-
Aminzadeh A, Roberts L, Young B, Chiang CI, Svalbe ID, Paganin DM, and Kingston AM
- Abstract
A set of non-configurable transversely-displaced masks has been designed and fabricated to generate high-quality X-ray illumination patterns for use in imaging techniques such as ghost imaging (GI), ghost projection, and speckle tracking. The designs include a range of random binary and orthogonal patterns, fabricated through a combination of photolithography and gold electroplating techniques. We experimentally demonstrated that a single wafer can be used as an illumination mask for GI, employing individual illumination patterns and also a mixture of patterns, using a laboratory X-ray source. The quality of the reconstructed X-ray ghost images has been characterized and evaluated through a range of metrics.
- Published
- 2023
- Full Text
- View/download PDF
Catalog
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.