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. COVID-19 and Housing Security in Remote Alaska Native Communities

   Nicewonger, Todd E. ; McNair, Lisa D. ; Fritz, Stacey ; Supple, Laura ; Criss-Carboy, Laura ( November 2023 , Virginia Tech Publishing)

   Nicewonger, Todd E. ; McNair, Lisa D. ; Fritz, Stacey (Ed.)

   https://pressbooks.lib.vt.edu/alaskanative/ At the start of the pandemic, the editors of this annotated bibliography initiated a remote (i.e., largely virtual) ethnographic research project that investigated how COVID-19 was impacting off-site modular construction practices in Alaska Native communities. Many of these communities are located off the road system and thus face not only dramatically higher costs but multiple logistical challenges in securing licensed tradesmen and construction crews and in shipping building supplies and equipment to their communities. These barriers, as well as the region’s long winters and short building seasons, complicate the construction of homes and related infrastructure projects. Historically, these communities have also grappled with inadequate housing, including severe overcrowding and poor-quality building stock that is rarely designed for northern Alaska’s climate (Marino 2015). Moreover, state and federal bureaucracies and their associated funding opportunities often further complicate home building by failing to accommodate the digital divide in rural Alaska and the cultural values and practices of Native communities.[1] It is not surprising, then, that as we were conducting fieldwork for this project, we began hearing stories about these issues and about how the restrictions caused by the pandemic were further exacerbating them. Amidst these stories, we learned about how modular home construction was being imagined as a possible means for addressing both the complications caused by the pandemic and the need for housing in the region (McKinstry 2021). As a result, we began to investigate how modular construction practices were figuring into emergent responses to housing needs in Alaska communities. We soon realized that we needed to broaden our focus to capture a variety of prefabricated building methods that are often colloquially or idiomatically referred to as “modular.” This included a range of prefabricated building systems (e.g., manufactured, volumetric modular, system-built, and Quonset huts and other reused military buildings[2]). Our further questions about prefabricated housing in the region became the basis for this annotated bibliography. Thus, while this bibliography is one of multiple methods used to investigate these issues, it played a significant role in guiding our research and helped us bring together the diverse perspectives we were hearing from our interviews with building experts in the region and the wider debates that were circulating in the media and, to a lesser degree, in academia. The actual research for each of three sections was carried out by graduate students Lauren Criss-Carboy and Laura Supple.[3] They worked with us to identify source materials and their hard work led to the team identifying three themes that cover intersecting topics related to housing security in Alaska during the pandemic. The source materials collected in these sections can be used in a variety of ways depending on what readers are interested in exploring, including insights into debates on housing security in the region as the pandemic was unfolding (2021-2022). The bibliography can also be used as a tool for thinking about the relational aspects of these themes or the diversity of ways in which information on housing was circulating during the pandemic (and the implications that may have had on community well-being and preparedness). That said, this bibliography is not a comprehensive analysis. Instead, by bringing these three sections together with one another to provide a snapshot of what was happening at that time, it provides a critical jumping off point for scholars working on these issues. The first section focuses on how modular housing figured into pandemic responses to housing needs. In exploring this issue, author Laura Supple attends to both state and national perspectives as part of a broader effort to situate Alaska issues with modular housing in relation to wider national trends. This led to the identification of multiple kinds of literature, ranging from published articles to publicly circulated memos, blog posts, and presentations. These materials are important source materials that will likely fade in the vastness of the Internet and thus may help provide researchers with specific insights into how off-site modular construction was used – and perhaps hyped – to address pandemic concerns over housing, which in turn may raise wider questions about how networks, institutions, and historical experiences with modular construction are organized and positioned to respond to major societal disruptions like the pandemic. As Supple pointed out, most of the material identified in this review speaks to national issues and only a scattering of examples was identified that reflect on the Alaskan context. The second section gathers a diverse set of communications exploring housing security and homelessness in the region. The lack of adequate, healthy housing in remote Alaska communities, often referred to as Alaska’s housing crisis, is well-documented and preceded the pandemic (Guy 2020). As the pandemic unfolded, journalists and other writers reported on the immense stress that was placed on already taxed housing resources in these communities (Smith 2020; Lerner 2021). The resulting picture led the editors to describe in their work how housing security in the region exists along a spectrum that includes poor quality housing as well as various forms of houselessness including, particularly relevant for the context, “hidden homelessness” (Hope 2020; Rogers 2020). The term houseless is a revised notion of homelessness because it captures a richer array of both permanent and temporary forms of housing precarity that people may experience in a region (Christensen et al. 2107). By identifying sources that reflect on the multiple forms of housing insecurity that people were facing, this section highlights the forms of disparity that complicated pandemic responses. Moreover, this section underscores ingenuity (Graham 2019; Smith 2020; Jason and Fashant 2021) that people on the ground used to address the needs of their communities. The third section provides a snapshot from the first year of the pandemic into how CARES Act funds were allocated to Native Alaska communities and used to address housing security. This subject was extremely complicated in Alaska due to the existence of for-profit Alaska Native Corporations and disputes over eligibility for the funds impacted disbursements nationwide. The resources in this section cover that dispute, impacts of the pandemic on housing security, and efforts to use the funds for housing as well as barriers Alaska communities faced trying to secure and use the funds. In summary, this annotated bibliography provides an overview of what was happening, in real time, during the pandemic around a specific topic: housing security in largely remote Alaska Native communities. The media used by housing specialists to communicate the issues discussed here are diverse, ranging from news reports to podcasts and from blogs to journal articles. This diversity speaks to the multiple ways in which information was circulating on housing at a time when the nightly news and radio broadcasts focused heavily on national and state health updates and policy developments. Finding these materials took time, and we share them here because they illustrate why attention to housing security issues is critical for addressing crises like the pandemic. For instance, one theme that emerged out of a recent National Science Foundation workshop on COVID research in the North NSF Conference[4] was that Indigenous communities are not only recovering from the pandemic but also evaluating lessons learned to better prepare for the next one, and resilience will depend significantly on more—and more adaptable—infrastructure and greater housing security. 

   more » « less
3. Provably efficient machine learning for quantum many-body problems

   https://doi.org/10.1126/science.abk3333  

   Huang, Hsin-Yuan ; Kueng, Richard ; Torlai, Giacomo ; Albert, Victor V. ; Preskill, John ( September 2022 , Science)

   INTRODUCTION Solving quantum many-body problems, such as finding ground states of quantum systems, has far-reaching consequences for physics, materials science, and chemistry. Classical computers have facilitated many profound advances in science and technology, but they often struggle to solve such problems. Scalable, fault-tolerant quantum computers will be able to solve a broad array of quantum problems but are unlikely to be available for years to come. Meanwhile, how can we best exploit our powerful classical computers to advance our understanding of complex quantum systems? Recently, classical machine learning (ML) techniques have been adapted to investigate problems in quantum many-body physics. So far, these approaches are mostly heuristic, reflecting the general paucity of rigorous theory in ML. Although they have been shown to be effective in some intermediate-size experiments, these methods are generally not backed by convincing theoretical arguments to ensure good performance. RATIONALE A central question is whether classical ML algorithms can provably outperform non-ML algorithms in challenging quantum many-body problems. We provide a concrete answer by devising and analyzing classical ML algorithms for predicting the properties of ground states of quantum systems. We prove that these ML algorithms can efficiently and accurately predict ground-state properties of gapped local Hamiltonians, after learning from data obtained by measuring other ground states in the same quantum phase of matter. Furthermore, under a widely accepted complexity-theoretic conjecture, we prove that no efficient classical algorithm that does not learn from data can achieve the same prediction guarantee. By generalizing from experimental data, ML algorithms can solve quantum many-body problems that could not be solved efficiently without access to experimental data. RESULTS We consider a family of gapped local quantum Hamiltonians, where the Hamiltonian H ( x ) depends smoothly on m parameters (denoted by x ). The ML algorithm learns from a set of training data consisting of sampled values of x , each accompanied by a classical representation of the ground state of H ( x ). These training data could be obtained from either classical simulations or quantum experiments. During the prediction phase, the ML algorithm predicts a classical representation of ground states for Hamiltonians different from those in the training data; ground-state properties can then be estimated using the predicted classical representation. Specifically, our classical ML algorithm predicts expectation values of products of local observables in the ground state, with a small error when averaged over the value of x . The run time of the algorithm and the amount of training data required both scale polynomially in m and linearly in the size of the quantum system. Our proof of this result builds on recent developments in quantum information theory, computational learning theory, and condensed matter theory. Furthermore, under the widely accepted conjecture that nondeterministic polynomial-time (NP)–complete problems cannot be solved in randomized polynomial time, we prove that no polynomial-time classical algorithm that does not learn from data can match the prediction performance achieved by the ML algorithm. In a related contribution using similar proof techniques, we show that classical ML algorithms can efficiently learn how to classify quantum phases of matter. In this scenario, the training data consist of classical representations of quantum states, where each state carries a label indicating whether it belongs to phase A or phase B . The ML algorithm then predicts the phase label for quantum states that were not encountered during training. The classical ML algorithm not only classifies phases accurately, but also constructs an explicit classifying function. Numerical experiments verify that our proposed ML algorithms work well in a variety of scenarios, including Rydberg atom systems, two-dimensional random Heisenberg models, symmetry-protected topological phases, and topologically ordered phases. CONCLUSION We have rigorously established that classical ML algorithms, informed by data collected in physical experiments, can effectively address some quantum many-body problems. These rigorous results boost our hopes that classical ML trained on experimental data can solve practical problems in chemistry and materials science that would be too hard to solve using classical processing alone. Our arguments build on the concept of a succinct classical representation of quantum states derived from randomized Pauli measurements. Although some quantum devices lack the local control needed to perform such measurements, we expect that other classical representations could be exploited by classical ML with similarly powerful results. How can we make use of accessible measurement data to predict properties reliably? Answering such questions will expand the reach of near-term quantum platforms. Classical algorithms for quantum many-body problems. Classical ML algorithms learn from training data, obtained from either classical simulations or quantum experiments. Then, the ML algorithm produces a classical representation for the ground state of a physical system that was not encountered during training. Classical algorithms that do not learn from data may require substantially longer computation time to achieve the same task. 

   more » « less
4. State-Aware Meta-Evaluation of Evaluation Metrics in Interactive Information Retrieval

   https://doi.org/10.1145/3459637.3482190  

   Liu, Jiqun ; Yu, Ran ( October 2021 , 30th ACM International Conference on Information & Knowledge Management)

   In interactive IR (IIR), users often seek to achieve different goals (e.g. exploring a new topic, finding a specific known item) at different search iterations and thus may evaluate system performances differently. Without state-aware approach, it would be extremely difficult to simulate and achieve real-time adaptive search evaluation and recommendation. To address this gap, our work identifies users' task states from interactive search sessions and meta-evaluates a series of online and offline evaluation metrics under varying states based on a user study dataset consisting of 1548 unique query segments from 450 search sessions. Our results indicate that: 1) users' individual task states can be identified and predicted from search behaviors and implicit feedback; 2) the effectiveness of mainstream evaluation measures (measured based upon their respective correlations with user satisfaction) vary significantly across task states. This study demonstrates the implicit heterogeneity in user-oriented IR evaluation and connects studies on complex search tasks with evaluation techniques. It also informs future research on the design of state-specific, adaptive user models and evaluation metrics. 

   more » « less
5. Variations in stability revealed by temporal asymmetries in contraction of phase space flow

   https://doi.org/10.1038/s41598-021-84865-8  

   Williams, Zachary C. ; McNamara, Dylan E. ( March 2021 , Scientific Reports)

   Empirical diagnosis of stability has received considerable attention, often focused on variance metrics for early warning signals of abrupt system change or delicate techniques measuring Lyapunov spectra. The theoretical foundation for the popular early warning signal approach has been limited to relatively simple system changes such as bifurcating fixed points where variability is extrinsic to the steady state. We offer a novel measurement of stability that applies in wide ranging systems that contain variability in both internal steady state dynamics and in response to external perturbations. Utilizing connections between stability, dissipation, and phase space flow, we show that stability correlates with temporal asymmetry in a measure of phase space flow contraction. Our method is general as it reveals stability variation independent of assumptions about the nature of system variability or attractor shape. After showing efficacy in a variety of model systems, we apply our technique for measuring stability to monthly returns of the S&P 500 index in the time periods surrounding the global stock market crash of October 1987. Market stability is shown to be higher in the several years preceding and subsequent to the 1987 market crash. We anticipate our technique will have wide applicability in climate, ecological, financial, and social systems where stability is a pressing concern.

    

   more » « less