David Carlson

Yiling Liu, Juncheng Dong, Ziyang Jiang et al.

Unsupervised domain adaptation (UDA) is a technique used to transfer knowledge from a labeled source domain to a different but related unlabeled target domain. While many UDA methods have shown success in the past, they often assume that the source and target domains must have identical class label distributions, which can limit their effectiveness in real-world scenarios. To address this limitation, we propose a novel generalization bound that reweights source classification error by aligning source and target sub-domains. We prove that our proposed generalization bound is at least as strong as existing bounds under realistic assumptions, and we empirically show that it is much stronger on real-world data. We then propose an algorithm to minimize this novel generalization bound. We demonstrate by numerical experiments that this approach improves performance in shifted class distribution scenarios compared to state-of-the-art methods.

Ziyang Jiang, Yiling Liu, Michael H. Klein et al.

Causal mediation analysis (CMA) is a powerful method to dissect the total effect of a treatment into direct and mediated effects within the potential outcome framework. This is important in many scientific applications to identify the underlying mechanisms of a treatment effect. However, in many scientific applications the mediator is unobserved, but there may exist related measurements. For example, we may want to identify how changes in brain activity or structure mediate an antidepressant's effect on behavior, but we may only have access to electrophysiological or imaging brain measurements. To date, most CMA methods assume that the mediator is one-dimensional and observable, which oversimplifies such real-world scenarios. To overcome this limitation, we introduce a CMA framework that can handle complex and indirectly observed mediators based on the identifiable variational autoencoder (iVAE) architecture. We prove that the true joint distribution over observed and latent variables is identifiable with the proposed method. Additionally, our framework captures a disentangled representation of the indirectly observed mediator and yields accurate estimation of the direct and mediated effects in synthetic and semi-synthetic experiments, providing evidence of its potential utility in real-world applications.

Ziyang Jiang, Zhuoran Hou, Yiling Liu et al.

A number of methods have been proposed for causal effect estimation, yet few have demonstrated efficacy in handling data with complex structures, such as images. To fill this gap, we propose Causal Multi-task Deep Ensemble (CMDE), a novel framework that learns both shared and group-specific information from the study population. We provide proofs demonstrating equivalency of CDME to a multi-task Gaussian process (GP) with a coregionalization kernel a priori. Compared to multi-task GP, CMDE efficiently handles high-dimensional and multi-modal covariates and provides pointwise uncertainty estimates of causal effects. We evaluate our method across various types of datasets and tasks and find that CMDE outperforms state-of-the-art methods on a majority of these tasks.

Ian Hough, Loïc Argentier, Ziyang Jiang et al.

Oxidative potential (OP), which measures particulate matter's (PM) capacity to induce oxidative stress in the lungs, is increasingly recognized as an indicator of PM toxicity. Since OP is not routinely monitored, it can be challenging to estimate exposure and health impacts. Remote sensing data are commonly used to estimate PM mass concentration, but have never been used to estimate OP. In this study, we evaluate the potential of satellite images to estimate OP as measured by acellular ascorbic acid (OP AA) and dithiothreitol (OP DTT) assays of 24-hour PM10 sampled periodically over five years at three locations around Grenoble, France. We use a deep convolutional neural network to extract features of daily 3 m/pixel PlanetScope satellite images and train a multilayer perceptron to estimate OP at a 1 km spatial resolution based on the image features and common meteorological variables. The model captures more than half of the variation in OP AA and almost half of the variation in OP DTT (test set R2 = 0.62 and 0.48, respectively), with relative mean absolute error (MAE) of about 32%. Using only satellite images, the model still captures about half of the variation in OP AA and one third of the variation in OP DTT (test set R2 = 0.49 and 0.36, respectively) with relative MAE of about 37%. If confirmed in other areas, our approach could represent a low-cost method for expanding the temporal or spatial coverage of OP estimates.

Tianhui Zhou, William E. Carson, Michael Hunter Klein et al.

Observational studies are regarded as economic alternatives to randomized trials, often used in their stead to investigate and determine treatment efficacy. Due to lack of sample size, observational studies commonly combine data from multiple sources or different sites/centers. Despite the benefits of an increased sample size, a naive combination of multicenter data may result in incongruities stemming from center-specific protocols for generating cohorts or reactions towards treatments distinct to a given center, among other things. These issues arise in a variety of other contexts, including capturing a treatment effect related to an individual's unique biological characteristics. Existing methods for estimating heterogeneous treatment effects have not adequately addressed the multicenter context, but rather treat it simply as a means to obtain sufficient sample size. Additionally, previous approaches to estimating treatment effects do not straightforwardly generalize to the multicenter design, especially when required to provide treatment insights for patients from a new, unobserved center. To address these shortcomings, we propose Multiple Domain Causal Networks (MDCN), an approach that simultaneously strengthens the information sharing between similar centers while addressing the selection bias in treatment assignment through learning of a new feature embedding. In empirical evaluations, MDCN is consistently more accurate when estimating the heterogeneous treatment effect in new centers compared to benchmarks that adjust solely based on treatment imbalance or general center differences. Finally, we justify our approach by providing theoretical analyses that demonstrate that MDCN improves on the generalization bound of the new, unobserved target center.

Ziyang Jiang, Tongshu Zheng, Yiling Liu et al.

It is challenging to guide neural network (NN) learning with prior knowledge. In contrast, many known properties, such as spatial smoothness or seasonality, are straightforward to model by choosing an appropriate kernel in a Gaussian process (GP). Many deep learning applications could be enhanced by modeling such known properties. For example, convolutional neural networks (CNNs) are frequently used in remote sensing, which is subject to strong seasonal effects. We propose to blend the strengths of deep learning and the clear modeling capabilities of GPs by using a composite kernel that combines a kernel implicitly defined by a neural network with a second kernel function chosen to model known properties (e.g., seasonality). We implement this idea by combining a deep network and an efficient mapping based on the Nystrom approximation, which we call Implicit Composite Kernel (ICK). We then adopt a sample-then-optimize approach to approximate the full GP posterior distribution. We demonstrate that ICK has superior performance and flexibility on both synthetic and real-world data sets. We believe that ICK framework can be used to include prior information into neural networks in many applications.

Karthikeyan K, Raghuveer Thirukovalluru, David Carlson

Clinical notes contain valuable, context-rich information, but their unstructured format introduces several challenges, including unintended biases (e.g., gender or racial bias), and poor generalization across clinical settings (e.g., models trained on one EHR system may perform poorly on another due to format differences) and poor interpretability. To address these issues, we present ClinStructor, a pipeline that leverages large language models (LLMs) to convert clinical free-text into structured, task-specific question-answer pairs prior to predictive modeling. Our method substantially enhances transparency and controllability and only leads to a modest reduction in predictive performance (a 2-3% drop in AUC), compared to direct fine-tuning, on the ICU mortality prediction task. ClinStructor lays a strong foundation for building reliable, interpretable, and generalizable machine learning models in clinical environments.

Karthikeyan K, Raghuveer Thirukovalluru, David Carlson

Large Language Models have advanced clinical text classification, but their opaque predictions remain a critical barrier to practical adoption in research and clinical settings where investigators and physicians need to understand which parts of a patient's record drive risk signals. To address this challenge, we introduce \textbf{CALM}, short for \textbf{Classification with Additive Large Language Models}, an interpretable framework for semi-structured text where inputs are composed of semantically meaningful components, such as sections of an admission note or question-answer fields from an intake form. CALM predicts outcomes as the additive sum of each component's contribution, making these contributions part of the forward computation itself and enabling faithful explanations at both the patient and population level. The additive structure also enables clear visualizations, such as component-level risk curves similar to those used in generalized additive models, making the learned relationships easier to inspect and communicate. Although CALM expects semi-structured inputs, many clinical documents already have this form, and similar structure can often be automatically extracted from free-text notes. CALM achieves performance comparable to conventional LLM classifiers while improving trust, supporting quality-assurance checks, and revealing clinically meaningful patterns during model development and auditing.

William E. Carson, Austin Talbot, David Carlson

Deep autoencoders are often extended with a supervised or adversarial loss to learn latent representations with desirable properties, such as greater predictivity of labels and outcomes or fairness with respects to a sensitive variable. Despite the ubiquity of supervised and adversarial deep latent factor models, these methods should demonstrate improvement over simpler linear approaches to be preferred in practice. This necessitates a reproducible linear analog that still adheres to an augmenting supervised or adversarial objective. We address this methodological gap by presenting methods that augment the principal component analysis (PCA) objective with either a supervised or an adversarial objective and provide analytic and reproducible solutions. We implement these methods in an open-source Python package, AugmentedPCA, that can produce excellent real-world baselines. We demonstrate the utility of these factor models on an open-source, RNA-seq cancer gene expression dataset, showing that augmenting with a supervised objective results in improved downstream classification performance, produces principal components with greater class fidelity, and facilitates identification of genes aligned with the principal axes of data variance with implications to development of specific types of cancer.

Zachary C. Brown, David Carlson

The field of hypothesis generation promises to reduce costs in neuroscience by narrowing the range of interventional studies needed to study various phenomena. Existing machine learning methods can generate scientific hypotheses from complex datasets, but many approaches assume causal relationships are static over time, limiting their applicability to systems with dynamic, state-dependent behavior, such as the brain. While some techniques attempt dynamic causal discovery through factor models, they often restrict relationships to linear patterns or impose other simplifying assumptions. We propose a novel method that models dynamic graphs as a conditionally weighted superposition of static graphs, where each static graph can capture nonlinear relationships. This approach enables the detection of complex, time-varying interactions between variables beyond linear limitations. Our method improves f1-scores of predicted dynamic causal patterns by roughly 22-28% on average over baselines in some of our experiments, with some improvements reaching well over 60%. A case study on real brain data demonstrates our method's ability to uncover relationships linked to specific behavioral states, offering valuable insights into neural dynamics.

Juncheng Dong, Yiling Liu, Ahmed Aloui et al.

Large language models (LLMs) have recently demonstrated impressive capabilities across a range of reasoning and generation tasks. However, research studies have shown that LLMs lack the ability to identify causal relationships, a fundamental cornerstone of human intelligence. We first conduct an exploratory investigation of LLMs' behavior when asked to perform a causal-discovery task and find that they mostly rely on the semantic meaning of variable names, ignoring the observation data. This is unsurprising, given that LLMs were never trained to process structural datasets. To first tackle this challenge, we prompt the LLMs with the outputs of established causal discovery algorithms designed for observational datasets. These algorithm outputs effectively serve as the sufficient statistics of the observation data. However, quite surprisingly, we find that prompting the LLMs with these sufficient statistics decreases the LLMs' performance in causal discovery. To address this current limitation, we propose CARE, a framework that enhances LLMs' causal-reasoning ability by teaching them to effectively utilize the outputs of established causal-discovery algorithms through supervised fine-tuning. Experimental results show that a finetuned Qwen2.5-1.5B model produced by CARE significantly outperforms both traditional causal-discovery algorithms and state-of-the-art LLMs with over a thousand times more parameters, demonstrating effective utilization of its own knowledge and the external algorithmic clues.

Yiling Liu, Juncheng Dong, Chen Fu et al.

Estimating counterfactual outcomes from time-series observations is crucial for effective decision-making, e.g. when to administer a life-saving treatment, yet remains significantly challenging because (i) the counterfactual trajectory is never observed and (ii) confounders evolve with time and distort estimation at every step. To address these challenges, we propose a novel framework that synergistically integrates two complementary approaches: Sub-treatment Group Alignment (SGA) and Random Temporal Masking (RTM). Instead of the coarse practice of aligning marginal distributions of the treatments in latent space, SGA uses iterative treatment-agnostic clustering to identify fine-grained sub-treatment groups. Aligning these fine-grained groups achieves improved distributional matching, thus leading to more effective deconfounding. We theoretically demonstrate that SGA optimizes a tighter upper bound on counterfactual risk and empirically verify its deconfounding efficacy. RTM promotes temporal generalization by randomly replacing input covariates with Gaussian noises during training. This encourages the model to rely less on potentially noisy or spuriously correlated covariates at the current step and more on stable historical patterns, thereby improving its ability to generalize across time and better preserve underlying causal relationships. Our experiments demonstrate that while applying SGA and RTM individually improves counterfactual outcome estimation, their synergistic combination consistently achieves state-of-the-art performance. This success comes from their distinct yet complementary roles: RTM enhances temporal generalization and robustness across time steps, while SGA improves deconfounding at each specific time point.

Ziyang Jiang, Zach Calhoun, Yiling Liu et al.

Causal reasoning is often challenging with spatial data, particularly when handling high-dimensional inputs. To address this, we propose a neural network (NN) based framework integrated with an approximate Gaussian process to manage spatial interference and unobserved confounding. Additionally, we adopt a generalized propensity-score-based approach to address partially observed outcomes when estimating causal effects with continuous treatments. We evaluate our framework using synthetic, semi-synthetic, and real-world data inferred from satellite imagery. Our results demonstrate that NN-based models significantly outperform linear spatial regression models in estimating causal effects. Furthermore, in real-world case studies, NN-based models offer more reasonable predictions of causal effects, facilitating decision-making in relevant applications.

Lei Duan, Ziyang Jiang, David Carlson

Fusing abundant satellite data with sparse ground measurements constitutes a major challenge in climate modeling. To address this, we propose a strategy to augment the training dataset by introducing unlabeled satellite images paired with pseudo-labels generated through a spatial interpolation technique known as ordinary kriging, thereby making full use of the available satellite data resources. We show that the proposed data augmentation strategy helps enhance the performance of the state-of-the-art convolutional neural network-random forest (CNN-RF) model by a reasonable amount, resulting in a noteworthy improvement in spatial correlation and a reduction in prediction error.

Tianhui Zhou, William E Carson, David Carlson

Traditional causal inference approaches leverage observational study data to estimate the difference in observed and unobserved outcomes for a potential treatment, known as the Conditional Average Treatment Effect (CATE). However, CATE corresponds to the comparison on the first moment alone, and as such may be insufficient in reflecting the full picture of treatment effects. As an alternative, estimating the full potential outcome distributions could provide greater insights. However, existing methods for estimating treatment effect potential outcome distributions often impose restrictive or simplistic assumptions about these distributions. Here, we propose Collaborating Causal Networks (CCN), a novel methodology which goes beyond the estimation of CATE alone by learning the full potential outcome distributions. Estimation of outcome distributions via the CCN framework does not require restrictive assumptions of the underlying data generating process. Additionally, CCN facilitates estimation of the utility of each possible treatment and permits individual-specific variation through utility functions. CCN not only extends outcome estimation beyond traditional risk difference, but also enables a more comprehensive decision-making process through definition of flexible comparisons. Under assumptions commonly made in the causal literature, we show that CCN learns distributions that asymptotically capture the true potential outcome distributions. Furthermore, we propose an adjustment approach that is empirically effective in alleviating sample imbalance between treatment groups in observational data. Finally, we evaluate the performance of CCN in multiple synthetic and semi-synthetic experiments. We demonstrate that CCN learns improved distribution estimates compared to existing Bayesian and deep generative methods as well as improved decisions with respects to a variety of utility functions.

William E. Carson, Dmitry Isaev, Samatha Major et al.

Discovering reliable measures that inform on autism spectrum disorder (ASD) diagnosis is critical for providing appropriate and timely treatment for this neurodevelopmental disorder. In this work we present applications of adversarial linear factor models in the creation of improved biomarkers for ASD diagnosis. First, we demonstrate that an adversarial linear factor model can be used to remove confounding information from our biomarkers, ensuring that they contain only pertinent information on ASD. Second, we show this same model can be used to learn a disentangled representation of multimodal biomarkers that results in an increase in predictive performance. These results demonstrate that adversarial methods can address both biomarker confounds and improve biomarker predictive performance.

Liyun Tu, Austin Talbot, Neil Gallagher et al.

Probabilistic generative models are attractive for scientific modeling because their inferred parameters can be used to generate hypotheses and design experiments. This requires that the learned model provide an accurate representation of the input data and yield a latent space that effectively predicts outcomes relevant to the scientific question. Supervised Variational Autoencoders (SVAEs) have previously been used for this purpose, where a carefully designed decoder can be used as an interpretable generative model while the supervised objective ensures a predictive latent representation. Unfortunately, the supervised objective forces the encoder to learn a biased approximation to the generative posterior distribution, which renders the generative parameters unreliable when used in scientific models. This issue has remained undetected as reconstruction losses commonly used to evaluate model performance do not detect bias in the encoder. We address this previously-unreported issue by developing a second order supervision framework (SOS-VAE) that influences the decoder to induce a predictive latent representation. This ensures that the associated encoder maintains a reliable generative interpretation. We extend this technique to allow the user to trade-off some bias in the generative parameters for improved predictive performance, acting as an intermediate option between SVAEs and our new SOS-VAE. We also use this methodology to address missing data issues that often arise when combining recordings from multiple scientific experiments. We demonstrate the effectiveness of these developments using synthetic data and electrophysiological recordings with an emphasis on how our learned representations can be used to design scientific experiments.

Austin Talbot, David Dunson, Kafui Dzirasa et al.

Targeted stimulation of the brain has the potential to treat mental illnesses. We propose an approach to help design the stimulation protocol by identifying electrical dynamics across many brain regions that relate to illness states. We model multi-region electrical activity as a superposition of activity from latent networks, where the weights on the latent networks relate to an outcome of interest. In order to improve on drawbacks of latent factor modeling in this context, we focus on supervised autoencoders (SAEs), which can improve predictive performance while maintaining a generative model. We explain why SAEs yield improved predictions, describe the distributional assumptions under which SAEs are an appropriate modeling choice, and provide modeling constraints to ensure biological relevance of the learned network. We use the analysis strategy to find a network associated with stress that characterizes a genotype associated with bipolar disorder. This discovered network aligns with a previously used stimulation technique, providing experimental validation of our approach.

Tianhui Zhou, Yitong Li, Yuan Wu et al.

Effective decision making requires understanding the uncertainty inherent in a prediction. In regression, this uncertainty can be estimated by a variety of methods; however, many of these methods are laborious to tune, generate overconfident uncertainty intervals, or lack sharpness (give imprecise intervals). We address these challenges by proposing a novel method to capture predictive distributions in regression by defining two neural networks with two distinct loss functions. Specifically, one network approximates the cumulative distribution function, and the second network approximates its inverse. We refer to this method as Collaborating Networks (CN). Theoretical analysis demonstrates that a fixed point of the optimization is at the idealized solution, and that the method is asymptotically consistent to the ground truth distribution. Empirically, learning is straightforward and robust. We benchmark CN against several common approaches on two synthetic and six real-world datasets, including forecasting A1c values in diabetic patients from electronic health records, where uncertainty is critical. In the synthetic data, the proposed approach essentially matches ground truth. In the real-world datasets, CN improves results on many performance metrics, including log-likelihood estimates, mean absolute errors, coverage estimates, and prediction interval widths.

Pengyu Cheng, Yitong Li, Xinyuan Zhang et al.

Textual network embedding aims to learn low-dimensional representations of text-annotated nodes in a graph. Prior work in this area has typically focused on fixed graph structures; however, real-world networks are often dynamic. We address this challenge with a novel end-to-end node-embedding model, called Dynamic Embedding for Textual Networks with a Gaussian Process (DetGP). After training, DetGP can be applied efficiently to dynamic graphs without re-training or backpropagation. The learned representation of each node is a combination of textual and structural embeddings. Because the structure is allowed to be dynamic, our method uses the Gaussian process to take advantage of its non-parametric properties. To use both local and global graph structures, diffusion is used to model multiple hops between neighbors. The relative importance of global versus local structure for the embeddings is learned automatically. With the non-parametric nature of the Gaussian process, updating the embeddings for a changed graph structure requires only a forward pass through the learned model. Considering link prediction and node classification, experiments demonstrate the empirical effectiveness of our method compared to baseline approaches. We further show that DetGP can be straightforwardly and efficiently applied to dynamic textual networks.

Cynthia Rudin, David Carlson

Despite the widespread usage of machine learning throughout organizations, there are some key principles that are commonly missed. In particular: 1) There are at least four main families for supervised learning: logical modeling methods, linear combination methods, case-based reasoning methods, and iterative summarization methods. 2) For many application domains, almost all machine learning methods perform similarly (with some caveats). Deep learning methods, which are the leading technique for computer vision problems, do not maintain an edge over other methods for most problems (and there are reasons why). 3) Neural networks are hard to train and weird stuff often happens when you try to train them. 4) If you don't use an interpretable model, you can make bad mistakes. 5) Explanations can be misleading and you can't trust them. 6) You can pretty much always find an accurate-yet-interpretable model, even for deep neural networks. 7) Special properties such as decision making or robustness must be built in, they don't happen on their own. 8) Causal inference is different than prediction (correlation is not causation). 9) There is a method to the madness of deep neural architectures, but not always. 10) It is a myth that artificial intelligence can do anything.

Yitong Li, Zhe Gan, Yelong Shen et al.

We propose a new task, called Story Visualization. Given a multi-sentence paragraph, the story is visualized by generating a sequence of images, one for each sentence. In contrast to video generation, story visualization focuses less on the continuity in generated images (frames), but more on the global consistency across dynamic scenes and characters -- a challenge that has not been addressed by any single-image or video generation methods. We therefore propose a new story-to-image-sequence generation model, StoryGAN, based on the sequential conditional GAN framework. Our model is unique in that it consists of a deep Context Encoder that dynamically tracks the story flow, and two discriminators at the story and image levels, to enhance the image quality and the consistency of the generated sequences. To evaluate the model, we modified existing datasets to create the CLEVR-SV and Pororo-SV datasets. Empirically, StoryGAN outperforms state-of-the-art models in image quality, contextual consistency metrics, and human evaluation.

Yitong Li, Martin Renqiang Min, Dinghan Shen et al.

Generating videos from text has proven to be a significant challenge for existing generative models. We tackle this problem by training a conditional generative model to extract both static and dynamic information from text. This is manifested in a hybrid framework, employing a Variational Autoencoder (VAE) and a Generative Adversarial Network (GAN). The static features, called "gist," are used to sketch text-conditioned background color and object layout structure. Dynamic features are considered by transforming input text into an image filter. To obtain a large amount of data for training the deep-learning model, we develop a method to automatically create a matched text-video corpus from publicly available online videos. Experimental results show that the proposed framework generates plausible and diverse videos, while accurately reflecting the input text information. It significantly outperforms baseline models that directly adapt text-to-image generation procedures to produce videos. Performance is evaluated both visually and by adapting the inception score used to evaluate image generation in GANs.

Ari Pakman, Dar Gilboa, David Carlson et al.

We introduce a novel stochastic version of the non-reversible, rejection-free Bouncy Particle Sampler (BPS), a Markov process whose sample trajectories are piecewise linear. The algorithm is based on simulating first arrival times in a doubly stochastic Poisson process using the thinning method, and allows efficient sampling of Bayesian posteriors in big datasets. We prove that in the BPS no bias is introduced by noisy evaluations of the log-likelihood gradient. On the other hand, we argue that efficiency considerations favor a small, controllable bias in the construction of the thinning proposals, in exchange for faster mixing. We introduce a simple regression-based proposal intensity for the thinning method that controls this trade-off. We illustrate the algorithm in several examples in which it outperforms both unbiased, but slowly mixing stochastic versions of BPS, as well as biased stochastic gradient-based samplers.

David Carlson, Patrick Stinson, Ari Pakman et al.

Partition functions of probability distributions are important quantities for model evaluation and comparisons. We present a new method to compute partition functions of complex and multimodal distributions. Such distributions are often sampled using simulated tempering, which augments the target space with an auxiliary inverse temperature variable. Our method exploits the multinomial probability law of the inverse temperatures, and provides estimates of the partition function in terms of a simple quotient of Rao-Blackwellized marginal inverse temperature probability estimates, which are updated while sampling. We show that the method has interesting connections with several alternative popular methods, and offers some significant advantages. In particular, we empirically find that the new method provides more accurate estimates than Annealed Importance Sampling when calculating partition functions of large Restricted Boltzmann Machines (RBM); moreover, the method is sufficiently accurate to track training and validation log-likelihoods during learning of RBMs, at minimal computational cost.

Changyou Chen, David Carlson, Zhe Gan et al.

Stochastic gradient Markov chain Monte Carlo (SG-MCMC) methods are Bayesian analogs to popular stochastic optimization methods; however, this connection is not well studied. We explore this relationship by applying simulated annealing to an SGMCMC algorithm. Furthermore, we extend recent SG-MCMC methods with two key components: i) adaptive preconditioners (as in ADAgrad or RMSprop), and ii) adaptive element-wise momentum weights. The zero-temperature limit gives a novel stochastic optimization method with adaptive element-wise momentum weights, while conventional optimization methods only have a shared, static momentum weight. Under certain assumptions, our theoretical analysis suggests the proposed simulated annealing approach converges close to the global optima. Experiments on several deep neural network models show state-of-the-art results compared to related stochastic optimization algorithms.

Chunyuan Li, Changyou Chen, David Carlson et al.

Effective training of deep neural networks suffers from two main issues. The first is that the parameter spaces of these models exhibit pathological curvature. Recent methods address this problem by using adaptive preconditioning for Stochastic Gradient Descent (SGD). These methods improve convergence by adapting to the local geometry of parameter space. A second issue is overfitting, which is typically addressed by early stopping. However, recent work has demonstrated that Bayesian model averaging mitigates this problem. The posterior can be sampled by using Stochastic Gradient Langevin Dynamics (SGLD). However, the rapidly changing curvature renders default SGLD methods inefficient. Here, we propose combining adaptive preconditioners with SGLD. In support of this idea, we give theoretical properties on asymptotic convergence and predictive risk. We also provide empirical results for Logistic Regression, Feedforward Neural Nets, and Convolutional Neural Nets, demonstrating that our preconditioned SGLD method gives state-of-the-art performance on these models.

Josh Merel, David Carlson, Liam Paninski et al.

Neuroprosthetic brain-computer interfaces function via an algorithm which decodes neural activity of the user into movements of an end effector, such as a cursor or robotic arm. In practice, the decoder is often learned by updating its parameters while the user performs a task. When the user's intention is not directly observable, recent methods have demonstrated value in training the decoder against a surrogate for the user's intended movement. We describe how training a decoder in this way is a novel variant of an imitation learning problem, where an oracle or expert is employed for supervised training in lieu of direct observations, which are not available. Specifically, we describe how a generic imitation learning meta-algorithm, dataset aggregation (DAgger, [1]), can be adapted to train a generic brain-computer interface. By deriving existing learning algorithms for brain-computer interfaces in this framework, we provide a novel analysis of regret (an important metric of learning efficacy) for brain-computer interfaces. This analysis allows us to characterize the space of algorithmic variants and bounds on their regret rates. Existing approaches for decoder learning have been performed in the cursor control setting, but the available design principles for these decoders are such that it has been impossible to scale them to naturalistic settings. Leveraging our findings, we then offer an algorithm that combines imitation learning with optimal control, which should allow for training of arbitrary effectors for which optimal control can generate goal-oriented control. We demonstrate this novel and general BCI algorithm with simulated neuroprosthetic control of a 26 degree-of-freedom model of an arm, a sophisticated and realistic end effector.

Zhe Gan, Chunyuan Li, Ricardo Henao et al.

Deep dynamic generative models are developed to learn sequential dependencies in time-series data. The multi-layered model is designed by constructing a hierarchy of temporal sigmoid belief networks (TSBNs), defined as a sequential stack of sigmoid belief networks (SBNs). Each SBN has a contextual hidden state, inherited from the previous SBNs in the sequence, and is used to regulate its hidden bias. Scalable learning and inference algorithms are derived by introducing a recognition model that yields fast sampling from the variational posterior. This recognition model is trained jointly with the generative model, by maximizing its variational lower bound on the log-likelihood. Experimental results on bouncing balls, polyphonic music, motion capture, and text streams show that the proposed approach achieves state-of-the-art predictive performance, and has the capacity to synthesize various sequences.