skip to main content

Title: A Bayesian experimental autonomous researcher for mechanical design
While additive manufacturing (AM) has facilitated the production of complex structures, it has also highlighted the immense challenge inherent in identifying the optimum AM structure for a given application. Numerical methods are important tools for optimization, but experiment remains the gold standard for studying nonlinear, but critical, mechanical properties such as toughness. To address the vastness of AM design space and the need for experiment, we develop a Bayesian experimental autonomous researcher (BEAR) that combines Bayesian optimization and high-throughput automated experimentation. In addition to rapidly performing experiments, the BEAR leverages iterative experimentation by selecting experiments based on all available results. Using the BEAR, we explore the toughness of a parametric family of structures and observe an almost 60-fold reduction in the number of experiments needed to identify high-performing structures relative to a grid-based search. These results show the value of machine learning in experimental fields where data are sparse.
Authors:
; ; ; ; ; ; ;
Award ID(s):
1661412
Publication Date:
NSF-PAR ID:
10186120
Journal Name:
Science Advances
Volume:
6
Issue:
15
Page Range or eLocation-ID:
eaaz1708
ISSN:
2375-2548
Sponsoring Org:
National Science Foundation
More Like this
  1. LEGOs are one of the most popular toys and are known to be useful as instructional tools in STEM education. In this work, we used LEGO structures to demonstrate the energetic size effect on structural strength. Many material's fexural strength decreases with increasing structural size. We seek to demonstrate this effect in LEGO beams. Fracture experiments were performed using 3-point bend beams built of 2 X 4 LEGO blocks in a periodic staggered arrangement. LEGO wheels were used as rollers on either ends of the specimens which were weight compensated by adding counterweights. [1] Specimens were loaded by hanging weights at their midspan and the maximum sustained load was recorded. Specimens with a built-in defect (crack) of half specimen height were considered. Beam height was varied from two to 32 LEGO blocks while keeping the in-plane aspect ratio constant. The specimen thickness was kept constant at one LEGO block. Slow-motion videos and sound recordings of fractures were captured to determine how the fracture originated and propagated through the specimen. Flexural stress was calculated based on nominal specimen dimensions and fracture toughness was calculated following ASTM E-399 standard expressions from Srawley (1976). [2] The results demonstrate that the LEGO beams indeedmore »exhibit a size effect on strength. For smaller beams the Uexural strength is higher than for larger beams. The dependence of strength on size is similar to that of Bažant’s size effect law [3] . Initiation of failure occurs consistently at the built-in defect. The staggered arrangement causes persistent crack branching which is more pronounced in larger specimens. The results also show that the apparent fracture toughness increases as the specimen size decreases. Further ongoing investigations consider the effects of the initial crack length on the size effect and the fracture response. The present work demonstrates that LEGO structures can serve as an instructional tool. We demonstrate principles of non-linear elastic fracture mechanics and highlight the importance of material microstructure (architecture) in fracture response. The experimental method is reproducible in a classroom setting without the need for complex facilities. This work was partially supported by the National Science Foundation (NSF) under the award #1662177 and the School of Mechanical Engineering at Purdue University. The authors acknowledge the support of Dr. Thomas Siegmund and Glynn Gallaway. [1] Khalilpour, S., BaniAsad, E. and Dehestani, M., 2019. A review on concrete fracture energy and effective parameters. Cement and Concrete research, 120, pp.294-321. [2] Srawley, J.E., 1976, January. Wide range stress intensity factor expressions for ASTM E 399 standard fracture toughness specimens. In Conf. of Am. Soc. for Testing and Mater., Committee E-24 (No. E-8654). [3] Bažant, Z.P., 1999. Size effect on structural strength: a review. Archive of applied Mechanics, 69(9), pp.703-725.« less
  2. Abstract

    Uncertainty quantification (UQ) in metal additive manufacturing (AM) has attracted tremendous interest in order to dramatically improve product reliability. Model-based UQ, which relies on the validity of a computational model, has been widely explored as a potential substitute for the time-consuming and expensive UQ solely based on experiments. However, its adoption in the practical AM process requires overcoming two main challenges: (1) the inaccurate knowledge of uncertainty sources and (2) the intrinsic uncertainty associated with the computational model. Here, we propose a data-driven framework to tackle these two challenges by combining high throughput physical/surrogate model simulations and the AM-Bench experimental data from the National Institute of Standards and Technology (NIST). We first construct a surrogate model, based on high throughput physical simulations, for predicting the three-dimensional (3D) melt pool geometry and its uncertainty with respect to AM parameters and uncertainty sources. We then employ a sequential Bayesian calibration method to perform experimental parameter calibration and model correction to significantly improve the validity of the 3D melt pool surrogate model. The application of the calibrated melt pool model to UQ of the porosity level, an important quality factor, of AM parts, demonstrates its potential use in AM quality control. Themore »proposed UQ framework can be generally applicable to different AM processes, representing a significant advance toward physics-based quality control of AM products.

    « less
  3. Obeid, I. ; Selesnik, I. ; Picone, J. (Ed.)
    The Neuronix high-performance computing cluster allows us to conduct extensive machine learning experiments on big data [1]. This heterogeneous cluster uses innovative scheduling technology, Slurm [2], that manages a network of CPUs and graphics processing units (GPUs). The GPU farm consists of a variety of processors ranging from low-end consumer grade devices such as the Nvidia GTX 970 to higher-end devices such as the GeForce RTX 2080. These GPUs are essential to our research since they allow extremely compute-intensive deep learning tasks to be executed on massive data resources such as the TUH EEG Corpus [2]. We use TensorFlow [3] as the core machine learning library for our deep learning systems, and routinely employ multiple GPUs to accelerate the training process. Reproducible results are essential to machine learning research. Reproducibility in this context means the ability to replicate an existing experiment – performance metrics such as error rates should be identical and floating-point calculations should match closely. Three examples of ways we typically expect an experiment to be replicable are: (1) The same job run on the same processor should produce the same results each time it is run. (2) A job run on a CPU and GPU should producemore »identical results. (3) A job should produce comparable results if the data is presented in a different order. System optimization requires an ability to directly compare error rates for algorithms evaluated under comparable operating conditions. However, it is a difficult task to exactly reproduce the results for large, complex deep learning systems that often require more than a trillion calculations per experiment [5]. This is a fairly well-known issue and one we will explore in this poster. Researchers must be able to replicate results on a specific data set to establish the integrity of an implementation. They can then use that implementation as a baseline for comparison purposes. A lack of reproducibility makes it very difficult to debug algorithms and validate changes to the system. Equally important, since many results in deep learning research are dependent on the order in which the system is exposed to the data, the specific processors used, and even the order in which those processors are accessed, it becomes a challenging problem to compare two algorithms since each system must be individually optimized for a specific data set or processor. This is extremely time-consuming for algorithm research in which a single run often taxes a computing environment to its limits. Well-known techniques such as cross-validation [5,6] can be used to mitigate these effects, but this is also computationally expensive. These issues are further compounded by the fact that most deep learning algorithms are susceptible to the way computational noise propagates through the system. GPUs are particularly notorious for this because, in a clustered environment, it becomes more difficult to control which processors are used at various points in time. Another equally frustrating issue is that upgrades to the deep learning package, such as the transition from TensorFlow v1.9 to v1.13, can also result in large fluctuations in error rates when re-running the same experiment. Since TensorFlow is constantly updating functions to support GPU use, maintaining an historical archive of experimental results that can be used to calibrate algorithm research is quite a challenge. This makes it very difficult to optimize the system or select the best configurations. The overall impact of all of these issues described above is significant as error rates can fluctuate by as much as 25% due to these types of computational issues. Cross-validation is one technique used to mitigate this, but that is expensive since you need to do multiple runs over the data, which further taxes a computing infrastructure already running at max capacity. GPUs are preferred when training a large network since these systems train at least two orders of magnitude faster than CPUs [7]. Large-scale experiments are simply not feasible without using GPUs. However, there is a tradeoff to gain this performance. Since all our GPUs use the NVIDIA CUDA® Deep Neural Network library (cuDNN) [8], a GPU-accelerated library of primitives for deep neural networks, it adds an element of randomness into the experiment. When a GPU is used to train a network in TensorFlow, it automatically searches for a cuDNN implementation. NVIDIA’s cuDNN implementation provides algorithms that increase the performance and help the model train quicker, but they are non-deterministic algorithms [9,10]. Since our networks have many complex layers, there is no easy way to avoid this randomness. Instead of comparing each epoch, we compare the average performance of the experiment because it gives us a hint of how our model is performing per experiment, and if the changes we make are efficient. In this poster, we will discuss a variety of issues related to reproducibility and introduce ways we mitigate these effects. For example, TensorFlow uses a random number generator (RNG) which is not seeded by default. TensorFlow determines the initialization point and how certain functions execute using the RNG. The solution for this is seeding all the necessary components before training the model. This forces TensorFlow to use the same initialization point and sets how certain layers work (e.g., dropout layers). However, seeding all the RNGs will not guarantee a controlled experiment. Other variables can affect the outcome of the experiment such as training using GPUs, allowing multi-threading on CPUs, using certain layers, etc. To mitigate our problems with reproducibility, we first make sure that the data is processed in the same order during training. Therefore, we save the data from the last experiment and to make sure the newer experiment follows the same order. If we allow the data to be shuffled, it can affect the performance due to how the model was exposed to the data. We also specify the float data type to be 32-bit since Python defaults to 64-bit. We try to avoid using 64-bit precision because the numbers produced by a GPU can vary significantly depending on the GPU architecture [11-13]. Controlling precision somewhat reduces differences due to computational noise even though technically it increases the amount of computational noise. We are currently developing more advanced techniques for preserving the efficiency of our training process while also maintaining the ability to reproduce models. In our poster presentation we will demonstrate these issues using some novel visualization tools, present several examples of the extent to which these issues influence research results on electroencephalography (EEG) and digital pathology experiments and introduce new ways to manage such computational issues.« less
  4. Biocompatible molecules with electronic functionality provide a promising substrate for biocompatible electronic devices and electronic interfacing with biological systems. Synthetic oligopeptides composed of an aromatic π-core flanked by oligopeptide wings are a class of molecules that can self-assemble in aqueous environments into supramolecular nanoaggregates with emergent optical and electronic activity. We present an integrated computational–experimental pipeline employing all-atom molecular dynamics simulations and experimental UV-visible spectroscopy within an active learning workflow using deep representational learning and multi-objective and multi-fidelity Bayesian optimization to design π-conjugated peptides programmed to self-assemble into elongated pseudo-1D nanoaggregates with a high degree of H-type co-facial stacking of the π-cores. We consider as our design space the 694 982 unique π-conjugated peptides comprising a quaterthiophene π-core flanked by symmetric oligopeptide wings up to five amino acids in length. After sampling only 1181 molecules (∼0.17% of the design space) by computation and 28 (∼0.004%) by experiment, we identify and experimentally validate a diversity of previously unknown high-performing molecules and extract interpretable design rules linking peptide sequence to emergent supramolecular structure and properties.
  5. We consider the problem of inferring the conditional independence graph (CIG) of a high-dimensional stationary multivariate Gaussian time series. A sparse-group lasso-based frequency-domain formulation of the problem has been considered in the literature where the objective is to estimate the sparse inverse power spectral density (PSD) of the data via optimization of a sparse-group lasso based penalized log-likelihood cost function that is formulated in the frequency-domain. The CIG is then inferred from the estimated inverse PSD. Optimization in the previous approach was performed using an alternating minimization (AM) approach whose performance depends upon choice of a penalty parameter. In this paper we investigate an alternating direction method of multipliers (ADMM) approach for optimization to mitigate dependence on the penalty parameter. We also investigate selection of the tuning parameters based on Bayesian information criterion, and illustrate our approach using synthetic and real data. Comparisons with the "usual" i.i.d. modeling of time series for graph estimation are also provided.