More Like this

1. Issues in the Reproducibility of Deep Learning Results

   https://doi.org/10.1109/SPMB47826.2019.9037840  

   Jean-Paul, S. ; Elseify, T. ; Obeid, I. ; Picone, J. ( December 2019 , IEEE Signal Processing in Medicine and Biology Symposium (SPMB))

   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 produce 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. 

   more » « less
2. GAN-LSTM Predictor for Failure Prognostics of Rolling Element Bearings

   https://doi.org/10.1109/ICPHM51084.2021.9486650  

   Lu, Hao ; Barzegar, Vahid ; Nemani, Venkat P. ; Hu, Chao ; Laflamme, Simon ; Zimmerman, Andrew T. ( July 2021 , 2021 IEEE International Conference on Prognostics and Health Management (ICPHM))

   Failure prognostics is the process of predicting the remaining useful life (RUL) of machine components, which is vital for the predictive maintenance of industrial machinery. This paper presents a new deep learning approach for failure prognostics of rolling element bearings based on a Long Short-Term Memory (LSTM) predictor trained simultaneously within a Generative Adversarial Network (GAN) architecture. The LSTM predictor takes the current and past observations of a well-defined health index as an input, uses those to forecast the future degradation trajectory, and then derives the RUL. Our proposed approach has three unique features: (1) Defining the bearing failure threshold by adopting an International Organization of Standardization (ISO) standard, making the approach industry-relevant; (2) Employing a GAN-based data augmentation technique to improve the accuracy and robustness of RUL prediction in cases where the deep learning model has access to only a small amount of training data; (3) Integrating the training process of the LSTM predictor within the GAN architecture. A joint training approach is utilized to ensure that the LSTM predictor model learns both the original and artificially generated data to capture the degradation trajectories. We utilize a publicly available accelerated run-to-failure dataset of rolling element bearings to assess the performance of the proposed approach. Results of a five-fold cross-validation study show that the integration of the LSTM predictor with GAN helps to decrease the average RUL prediction error by 29% over a simple LSTM model without GAN implementation. 

   more » « less
3. Cross-frequency training with adversarial learning for radar micro-Doppler signature classification (Rising Researcher)

   https://doi.org/10.1117/12.2559155  

   Gurbuz, Sevgi Zubeyde ; Rahman, Mohammad M. ; Kurtoglu, Emre ; Macks, Trevor ; Fioranelli, Francesco ( May 2020 , Proc. SPIE 11408, Radar Sensor Technology XXIV)

   Deep neural networks have become increasingly popular in radar micro-Doppler classification; yet, a key challenge, which has limited potential gains, is the lack of large amounts of measured data that can facilitate the design of deeper networks with greater robustness and performance. Several approaches have been proposed in the literature to address this problem, such as unsupervised pre-training and transfer learning from optical imagery or synthetic RF data. This work investigates an alternative approach to training which involves exploitation of “datasets of opportunity” – micro-Doppler datasets collected using other RF sensors, which may be of a different frequency, bandwidth or waveform - for the purposes of training. Specifically, this work compares in detail the cross-frequency training degradation incurred for several different training approaches and deep neural network (DNN) architectures. Results show a 70% drop in classification accuracy when the RF sensors for pre-training, fine-tuning, and testing are different, and a 15% degradation when only the pre-training data is different, but the fine-tuning and test data are from the same sensor. By using generative adversarial networks (GANs), a large amount of synthetic data is generated for pre-training. Results show that cross-frequency performance degradation is reduced by 50% when kinematically-sifted GAN-synthesized signatures are used in pre-training. 

   more » « less
4. DeepMicroGen: a generative adversarial network-based method for longitudinal microbiome data imputation

   https://doi.org/10.1093/bioinformatics/btad286  

   Choi, Joung Min ; Ji, Ming ; Watson, Layne T. ; Zhang, Liqing ; Boeva, ed., Valentina ( April 2023 , Bioinformatics)

   The human microbiome, which is linked to various diseases by growing evidence, has a profound impact on human health. Since changes in the composition of the microbiome across time are associated with disease and clinical outcomes, microbiome analysis should be performed in a longitudinal study. However, due to limited sample sizes and differing numbers of timepoints for different subjects, a significant amount of data cannot be utilized, directly affecting the quality of analysis results. Deep generative models have been proposed to address this lack of data issue. Specifically, a generative adversarial network (GAN) has been successfully utilized for data augmentation to improve prediction tasks. Recent studies have also shown improved performance of GAN-based models for missing value imputation in a multivariate time series dataset compared with traditional imputation methods.

   This work proposes DeepMicroGen, a bidirectional recurrent neural network-based GAN model, trained on the temporal relationship between the observations, to impute the missing microbiome samples in longitudinal studies. DeepMicroGen outperforms standard baseline imputation methods, showing the lowest mean absolute error for both simulated and real datasets. Finally, the proposed model improved the predicted clinical outcome for allergies, by providing imputation for an incomplete longitudinal dataset used to train the classifier.

   DeepMicroGen is publicly available at https://github.com/joungmin-choi/DeepMicroGen.

    

   more » « less
5. Dynamic reliability management for multi-core processor based on deep reinforcement learning

   Z. Sun, H. Zhou ; Tan, Sheldon ( July 2019 , International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design (SMACD’19))

   In this paper, we propose a new dynamic reliability management (DRM) approach with deep reinforcement learning (DRL) for multi-core processors considering device reliability effects (hard error) and transient error of signal (soft error). The proposed method is based on a recently proposed physics-based three-phase electromigration model and an exponential soft error model that considers dynamic voltage and frequency scaling (DVFS) effects. Our work has been inspired by the recent advancements in DRL for various control and game applications. Compared with the traditional Q-learning based method, DRL has better scalability, lower memory and lower computational complexity. A large class of multi-threaded applications are used as the benchmark to validate and compare the proposed dynamic reliability management methods. Experimental results show that the proposed method can significantly reduces memory footprint and computational time compared to the traditional Q-learning based method. Furthermore, we show that the DRL-based DRM method can save 53.50% more energy than the Q-learning based method and 61.29% more than the simple DVFS based method. 

   more » « less