Computer-Aided-Diagnosis (CADx) systems assist radiologists with identifying and classifying potentially malignant pulmonary nodules on chest CT scans using morphology and texture-based (radiomic) features. However, radiomic features are sensitive to differences in acquisitions due to variations in dose levels and slice thickness. This study investigates the feasibility of generating a normalized scan from heterogeneous CT scans as input. We obtained projection data from 40 low-dose chest CT scans, simulating acquisitions at 10%, 25% and 50% dose and reconstructing the scans at 1.0mm and 2.0mm slice thickness. A 3D generative adversarial network (GAN) was used to simultaneously normalize reduced dose, thick slice (2.0mm) images to normal dose (100%), thinner slice (1.0mm) images. We evaluated the normalized image quality using peak signal-to-noise ratio (PSNR), structural similarity index (SSIM) and Learned Perceptual Image Patch Similarity (LPIPS). Our GAN improved perceptual similarity by 35%, compared to a baseline CNN method. Our analysis also shows that the GAN-based approach led to a significantly smaller error (p-value < 0.05) in nine studied radiomic features. These results indicated that GANs could be used to normalize heterogeneous CT images and reduce the variability in radiomic feature values.

T o encourage disentanglement, we devise a distance covariance based decorrelation regularization. Further, for the reconstruction step, our model leverages a soft target representation combined with the latent image code. By exploring the real-valued space of the soft target representation, we are able to synthesize novel images with the designated properties. T o improve the perceptual quality of images generated by autoencoder (AE)- based models, we extend the encoder-decoder architecture with the generative adversarial network (GAN) by collapsing the AE decoder and the GAN generator into one. We also design a classification based protocol to quantitatively evaluate the disentanglement strength of our model. Experimental results showcase the benefits of the proposed model.

Predicting response to neoadjuvant therapy is a vexing challenge in breast cancer. In this study, we evaluate the ability of deep learning to predict response to HER2-targeted neo-adjuvant chemotherapy (NAC) from pre-treatment dynamic contrast-enhanced (DCE) MRI acquired prior to treatment. In a retrospective study encompassing DCE-MRI data from a total of 157 HER2+ breast cancer patients from 5 institutions, we developed and validated a deep learning approach for predicting pathological complete response (pCR) to HER2-targeted NAC prior to treatment. 100 patients who received HER2-targeted neoadjuvant chemotherapy at a single institution were used to train (n=85) and tune (n=15) a convolutional neural network (CNN) to predict pCR. A multi-input CNN leveraging both pre-contrast and late post-contrast DCE-MRI acquisitions was identified to achieve optimal response prediction within the validation set (AUC=0.93). This model was then tested on two independent testing cohorts with pre-treatment DCE-MRI data. It achieved strong performance in a 28 patient testing set from a second institution (AUC=0.85, 95% CI 0.67-1.0, p=.0008) and a 29 patient multicenter trial including data from 3 additional institutions (AUC=0.77, 95% CI 0.58-0.97, p=0.006). Deep learning-based response prediction model was found to exceed a multivariable model incorporating predictive clinical variables (AUC < .65 in testing cohorts) and a model of semi-quantitative DCE-MRI pharmacokinetic measurements (AUC < .60 in testing cohorts). The results presented in this work across multiple sites suggest that with further validation deep learning could provide an effective and reliable tool to guide targeted therapy in breast cancer, thus reducing overtreatment among HER2+ patients.

We study empirical scaling laws for language model performance on the cross-entropy loss. The loss scales as a power-law with model size, dataset size, and the amount of compute used for training, with some trends spanning more than seven orders of magnitude. Other architectural details such as network width or depth have minimal effects within a wide range. Simple equations govern the dependence of overfitting on model/dataset size and the dependence of training speed on model size. These relationships allow us to determine the optimal allocation of a fixed compute budget. Larger models are significantly more sample-efficient, such that optimally compute-efficient training involves training very large models on a relatively modest amount of data and stopping significantly before convergence.

Autoencoder-based learning has emerged as a staple for disciplining representations in unsupervised and semi-supervised settings. This paper analyzes a framework for improving generalization in a purely supervised setting, where the target space is high-dimensional. We motivate and formalize the general framework of target-embedding autoencoders (TEA) for supervised prediction, learning intermediate latent representations jointly optimized to be both predictable from features as well as predictive of targets---encoding the prior that variations in targets are driven by a compact set of underlying factors. As our theoretical contribution, we provide a guarantee of generalization for linear TEAs by demonstrating uniform stability, interpreting the benefit of the auxiliary reconstruction task as a form of regularization. As our empirical contribution, we extend validation of this approach beyond existing static classification applications to multivariate sequence forecasting, verifying their advantage on both linear and nonlinear recurrent architectures---thereby underscoring the further generality of this framework beyond feedforward instantiations.

This paper presents several strategies that can improve neu - ral network-based predictive methods for MOOC student course trajectory modeling, applying multiple ideas previ - ously applied to tackle NLP (Natural Language Processing) tasks. In particular, this paper investigates LSTM network s enhanced with two forms of regularization, along with the more recently introduced Transformer architecture.

In this paper, we present a new approach to time series forecasting. Time series data are prevalent in many scientific and engineering disciplines. Time series forecasting is a crucial task in modeling time series data, and is an important area of machine learning. In this work we developed a novel method that employs Transformer-based machine learning models to forecast time series data. This approach works by leveraging self-attention mechanisms to learn complex patterns and dynamics from time series data. Moreover, it is a generic framework and can be applied to univariate and multivariate time series data, as well as time series embeddings. Using influenza-like illness (ILI) forecasting as a case study, we show that the forecasting results produced by our approach are favorably comparable to the state-of-the-art.

Graph generative models have been extensively studied in the data mining literature. While traditional techniques are based on generating structures that adhere to a pre-decided distribution, recent techniques have shifted towards learning this distribution directly from the data. While learning-based approaches have imparted significant improvement in quality, some limitations remain to be addressed. First, learning graph distributions introduces additional computational overhead, which limits their scalability to large graph databases. Second, many techniques only learn the structure and do not address the need to also learn node and edge labels, which encode important semantic information and influence the structure itself. Third, existing techniques often incorporate domain-specific rules and lack generalizability. Fourth, the experimentation of existing techniques is not comprehensive enough due to either using weak evaluation metrics or focusing primarily on synthetic or small datasets. In this work, we develop a domain-agnostic technique called GraphGen to overcome all of these limitations. GraphGen converts graphs to sequences using minimum DFS codes. Minimum DFS codes are canonical labels and capture the graph structure precisely along with the label information. The complex joint distributions between structure and semantic labels are learned through a novel LSTM architecture. Extensive experiments on million-sized, real graph datasets show GraphGen to be 4 times faster on average than state-of-the-art techniques while being significantly better in quality across a comprehensive set of 11 different metrics. Our code is released at https://github.com/idea-iitd/graphgen.

Optimizing fluid-dynamic performance is an important engineering task. Traditionally, experts design shapes based on empirical estimations and verify them through expensive experiments. This costly process, both in terms of time and space, may only explore a limited number of shapes and lead to sub-optimal designs. In this research, a test-proven deep learning architecture is applied to predict the performance under various restrictions and search for better shapes by optimizing the learned prediction function. The major challenge is the vast amount of data points Deep Neural Network (DNN) demands, which is improvident to simulate. To remedy this drawback, a Frequentist active learning is used to explore regions of the output space that DNN predicts promising. This operation reduces the number of data samples demanded from ~8000 to 625. The final stage, a user interface, made the model capable of optimizing with given user input of minimum area and viscosity. Flood fill is used to define a boundary area function so that the optimal shape does not bypass the minimum area. Stochastic Gradient Langevin Dynamics (SGLD) is employed to make sure the ultimate shape is optimized while circumventing the required area. Jointly, shapes with extremely low drags are found explored by a practical user interface with no human domain knowledge and modest computation overhead.

Uncertainty quantification for deep learning is a challenging open problem. Bayesian statistics offer a mathematically grounded framework to reason about uncertainties; however, approximate posteriors for modern neural networks still require prohibitive computational costs. We propose a family of algorithms which split the classification task into two stages: representation learning and uncertainty estimation. We compare four specific instances, where uncertainty estimation is performed via either an ensemble of Stochastic Gradient Descent or Stochastic Gradient Langevin Dynamics snapshots, an ensemble of bootstrapped logistic regressions, or via a number of Monte Carlo Dropout passes. We evaluate their performance in terms of \emph{selective} classification (risk-coverage), and their ability to detect out-of-distribution samples. Our experiments suggest there is limited value in adding multiple uncertainty layers to deep classifiers, and we observe that these simple methods strongly outperform a vanilla point-estimate SGD in some complex benchmarks like ImageNet.