Search

As artificial intelligence (AI) continues to advance, it demonstrates capabilities comparable to human intelligence, with significant potential to transform education and workforce development. This study evaluates OpenAI o1-preview's ability to perform higher-order cognitive tasks across 14 dimensions, including critical thinking, systems thinking, computational thinking, design thinking, metacognition, data literacy, creative thinking, abstract reasoning, quantitative reasoning, logical reasoning, analogical reasoning, and scientific reasoning. We used validated instruments like the Ennis-Weir Critical Thinking Essay Test and the Biological Systems Thinking Test to compare the o1-preview's performance with human performance systematically. Our findings reveal that o1-preview outperforms humans in most categories, achieving 150% better results in systems thinking, computational thinking, data literacy, creative thinking, scientific reasoning, and abstract reasoning. However, compared to humans, it underperforms by around 25% in logical reasoning, critical thinking, and quantitative reasoning. In analogical reasoning, both o1-preview and humans achieved perfect scores. Despite these strengths, the o1-preview shows limitations in abstract reasoning, where human psychology students outperform it, highlighting the continued importance of human oversight in tasks requiring high-level abstraction. These results have significant educational implications, suggesting a shift toward developing human skills that complement AI, such as creativity, abstract reasoning, and critical thinking. This study emphasizes the transformative potential of AI in education and calls for a recalibration of educational goals, teaching methods, and curricula to align with an AI-driven world., Comment: An assessment of OpenAI o1-Preview for Higher Order Thinking in Education

TeNPy (short for 'Tensor Network Python') is a python library for the simulation of strongly correlated quantum systems with tensor networks. The philosophy of this library is to achieve a balance of readability and usability for new-comers, while at the same time providing powerful algorithms for experts. The focus is on MPS algorithms for 1D and 2D lattices, such as DMRG ground state search, as well as dynamics using TEBD, TDVP, or MPO evolution. This article is a companion to the recent version 1.0 release of TeNPy and gives a brief overview of the package., Comment: v3: published version with small additional clarifications suggested by referees

Neutral atoms become strongly interacting when their electrons are excited to loosely bound Rydberg states. We investigate the strongly correlated quantum phases of matter that emerge in two-dimensional atom arrays where three Rydberg levels are used to encode an effective spin-1 degree of freedom. Dipolar exchange between such spin-1 Rydberg atoms naturally yields two distinct models: (i) a two-species hardcore boson model, and (ii) upon tuning near a F\"orster resonance, a dipolar spin-1 XY model. Through extensive, large-scale infinite density matrix renormalization group calculations, we provide a broad roadmap predicting the quantum phases that emerge from these models on a variety of lattice geometries: square, triangular, kagome, and ruby. We identify a wealth of correlated states, including lattice supersolids and simplex phases, all of which can be naturally realized in near-term experiments., Comment: 5 pages, 4 figures

The negatively charged boron vacancy ($\mathrm{V}_{\mathrm{B}}^-$) in hexagonal boron nitride (hBN) has garnered significant attention among defects in two-dimensional materials. This owes, in part, to its deterministic generation, well-characterized atomic structure, and optical polarizability at room temperature. We investigate the latter through extensive measurements probing both the ground and excited state polarization dynamics. We develop a semiclassical model based on these measurements that predicts a near-unity degree of spin polarization, surpassing other solid-state spin defects under ambient conditions. Building upon our model, we include the presence of nuclear spin degrees of freedom adjacent to the $\mathrm{V}_{\mathrm{B}}^-$ and perform a comprehensive set of Lindbladian numerics to investigate the hyperfine-induced polarization of the nuclear spins. Our simulations predict a number of important features that emerge as a function of magnetic field which are borne out by experiment.

We predict that the gapless $U(1)$ Dirac spin liquid naturally emerges in a two-dimensional array of quantum dipoles. In particular, we demonstrate that the dipolar XY model$\unicode{x2014}$realized in both Rydberg atom arrays and ultracold polar molecules$\unicode{x2014}$hosts a quantum spin liquid ground state on the kagome lattice. Large-scale density matrix renormalization group calculations indicate that this spin liquid exhibits signatures of gapless, linearly-dispersing spinons, consistent with the $U(1)$ Dirac spin liquid. We identify a route to adiabatic preparation via staggered on-site fields and demonstrate that this approach can prepare cold spin liquids within experimentally realistic time-scales. Finally, we propose a number of novel signatures of the Dirac spin liquid tailored to near-term quantum simulators, including termination-dependent edge modes and the Friedel response to a local perturbation., Comment: 5+13 pages, 4+10 figures

Computing is at a moment of profound opportunity. Emerging applications -- such as capable artificial intelligence, immersive virtual realities, and pervasive sensor systems -- drive unprecedented demand for computer. Despite recent advances toward net zero carbon emissions, the computing industry's gross energy usage continues to rise at an alarming rate, outpacing the growth of new energy installations and renewable energy deployments. A shift towards sustainability is needed to spark a transformation in how computer systems are manufactured, allocated, and consumed. Carbon Connect envisions coordinated research thrusts that produce design and management strategies for sustainable, next-generation computer systems. These strategies must flatten and then reverse growth trajectories for computing power and carbon for society's most rapidly growing applications such as artificial intelligence and virtual spaces. We will require accurate models for carbon accounting in computing technology. For embodied carbon, we must re-think conventional design strategies -- over-provisioned monolithic servers, frequent hardware refresh cycles, custom silicon -- and adopt life-cycle design strategies that more effectively reduce, reuse and recycle hardware at scale. For operational carbon, we must not only embrace renewable energy but also design systems to use that energy more efficiently. Finally, new hardware design and management strategies must be cognizant of economic policy and regulatory landscape, aligning private initiatives with societal goals. Many of these broader goals will require computer scientists to develop deep, enduring collaborations with researchers in economics, law, and industrial ecology to spark change in broader practice.

In the dynamic hospital setting, decision support can be a valuable tool for improving patient outcomes. Data-driven inference of future outcomes is challenging in this dynamic setting, where long sequences such as laboratory tests and medications are updated frequently. This is due in part to heterogeneity of data types and mixed-sequence types contained in variable length sequences. In this work we design a probabilistic unsupervised model for multiple arbitrary-length sequences contained in hospitalization Electronic Health Record (EHR) data. The model uses a latent variable structure and captures complex relationships between medications, diagnoses, laboratory tests, neurological assessments, and medications. It can be trained on original data, without requiring any lossy transformations or time binning. Inference algorithms are derived that use partial data to infer properties of the complete sequences, including their length and presence of specific values. We train this model on data from subjects receiving medical care in the Kaiser Permanente Northern California integrated healthcare delivery system. The results are evaluated against held-out data for predicting the length of sequences and presence of Intensive Care Unit (ICU) in hospitalization bed sequences. Our method outperforms a baseline approach, showing that in these experiments the trained model captures information in the sequences that is informative of their future values.

We propose training fitted Q-iteration with log-loss (FQI-log) for batch reinforcement learning (RL). We show that the number of samples needed to learn a near-optimal policy with FQI-log scales with the accumulated cost of the optimal policy, which is zero in problems where acting optimally achieves the goal and incurs no cost. In doing so, we provide a general framework for proving small-cost bounds, i.e. bounds that scale with the optimal achievable cost, in batch RL. Moreover, we empirically verify that FQI-log uses fewer samples than FQI trained with squared loss on problems where the optimal policy reliably achieves the goal.

At continuous phase transitions, quantum many-body systems exhibit scale-invariance and complex, emergent universal behavior. Most strikingly, at a quantum critical point, correlations decay as a power law, with exponents determined by a set of universal scaling dimensions. Experimentally probing such power-law correlations is extremely challenging, owing to the complex interplay between decoherence, the vanishing energy gap, and boundary effects. Here, we employ a Rydberg quantum simulator to adiabatically prepare critical ground states of both a one-dimensional ring and a two-dimensional square lattice. By accounting for and tuning the openness of our quantum system, which is well-captured by the introduction of a single phenomenological length scale, we are able to directly observe power-law correlations and extract the corresponding scaling dimensions. Moreover, in two dimensions, we observe a decoupling between phase transitions in the bulk and on the boundary, allowing us to identify two distinct boundary universality classes. Our work demonstrates that direct adiabatic preparation of critical states in quantum simulators can complement recent approaches to studying quantum criticality using the Kibble-Zurek mechanism or digital quantum circuits.

This work delves into the expanding role of large language models (LLMs) in generating artificial data. LLMs are increasingly employed to create a variety of outputs, including annotations, preferences, instruction prompts, simulated dialogues, and free text. As these forms of LLM-generated data often intersect in their application, they exert mutual influence on each other and raise significant concerns about the quality and diversity of the artificial data incorporated into training cycles, leading to an artificial data ecosystem. To the best of our knowledge, this is the first study to aggregate various types of LLM-generated text data, from more tightly constrained data like "task labels" to more lightly constrained "free-form text". We then stress test the quality and implications of LLM-generated artificial data, comparing it with human data across various existing benchmarks. Despite artificial data's capability to match human performance, this paper reveals significant hidden disparities, especially in complex tasks where LLMs often miss the nuanced understanding of intrinsic human-generated content. This study critically examines diverse LLM-generated data and emphasizes the need for ethical practices in data creation and when using LLMs. It highlights the LLMs' shortcomings in replicating human traits and behaviors, underscoring the importance of addressing biases and artifacts produced in LLM-generated content for future research and development. All data and code are available on our project page., Comment: Core Authors: Debarati Das, Karin De Langis, Anna Martin-Boyle, Jaehyung Kim, Minhwa Lee and Zae Myung Kim | Project lead : Debarati Das | PI : Dongyeop Kang

Reachable sets for a dynamical system describe collections of system states that can be reached in finite time, subject to system dynamics. They can be used to guarantee goal satisfaction in controller design or to verify that unsafe regions will be avoided. However, general-purpose methods for computing these sets suffer from the curse of dimensionality, which typically prohibits their use for systems with more than a small number of states, even if they are linear. In this paper, we demonstrate that viscosity supersolutions and subsolutions of a Hamilton-Jacobi-Bellman equation can be used to generate, respectively, under-approximating and over-approximating reachable sets for time-varying nonlinear systems. Based on this observation, we derive dynamics for a union and intersection of ellipsoidal sets that, respectively, under-approximate and over-approximate the reachable set for linear time-varying systems subject to an ellipsoidal input constraint and an ellipsoidal terminal (or initial) set. We demonstrate that the dynamics for these ellipsoids can be selected to ensure that their boundaries coincide with the boundary of the exact reachable set along a solution of the system. The ellipsoidal sets can be generated with polynomial computational complexity in the number of states, making our approximation scheme computationally tractable for continuous-time linear time-varying systems of relatively high dimension., Comment: 32 pages

Offline reinforcement learning algorithms often require careful hyperparameter tuning. Consequently, before deployment, we need to select amongst a set of candidate policies. As yet, however, there is little understanding about the fundamental limits of this offline policy selection (OPS) problem. In this work we aim to provide clarity on when sample efficient OPS is possible, primarily by connecting OPS to off-policy policy evaluation (OPE) and Bellman error (BE) estimation. We first show a hardness result, that in the worst case, OPS is just as hard as OPE, by proving a reduction of OPE to OPS. As a result, no OPS method can be more sample efficient than OPE in the worst case. We then propose a BE method for OPS, called Identifiable BE Selection (IBES), that has a straightforward method for selecting its own hyperparameters. We highlight that using IBES for OPS generally has more requirements than OPE methods, but if satisfied, can be more sample efficient. We conclude with an empirical study comparing OPE and IBES, and by showing the difficulty of OPS on an offline Atari benchmark dataset.

We use a Rydberg quantum simulator to demonstrate a new form of spectroscopy, called quench spectroscopy, which probes the low-energy excitations of a many-body system. We illustrate the method on a two-dimensional simulation of the spin-1/2 dipolar XY model. Through microscopic measurements of the spatial spin correlation dynamics following a quench, we extract the dispersion relation of the elementary excitations for both ferro- and anti-ferromagnetic couplings. We observe qualitatively different behaviors between the two cases that result from the long-range nature of the interactions, and the frustration inherent in the antiferromagnet. In particular, the ferromagnet exhibits elementary excitations behaving as linear spin waves. In the anti-ferromagnet, spin waves appear to decay, suggesting the presence of strong nonlinearities. Our demonstration highlights the importance of power-law interactions on the excitation spectrum of a many-body system., Comment: Main text 8 pages with 4 figures ; Supplemental Material 16 pages and 13 figures

Catastrophic interference is common in many network-based learning systems, and many proposals exist for mitigating it. Before overcoming interference we must understand it better. In this work, we provide a definition and novel measure of interference for value-based reinforcement learning methods such as Fitted Q-Iteration and DQN. We systematically evaluate our measure of interference, showing that it correlates with instability in control performance, across a variety of network architectures. Our new interference measure allows us to ask novel scientific questions about commonly used deep learning architectures and study learning algorithms which mitigate interference. Lastly, we outline a class of algorithms which we call online-aware that are designed to mitigate interference, and show they do reduce interference according to our measure and that they improve stability and performance in several classic control environments., Comment: Published at Conference on Lifelong Learning Agents (CoLLAs) 2023

In this work, we consider the off-policy policy evaluation problem for contextual bandits and finite horizon reinforcement learning in the nonstationary setting. Reusing old data is critical for policy evaluation, but existing estimators that reuse old data introduce large bias such that we can not obtain a valid confidence interval. Inspired from a related field called survey sampling, we introduce a variant of the doubly robust (DR) estimator, called the regression-assisted DR estimator, that can incorporate the past data without introducing a large bias. The estimator unifies several existing off-policy policy evaluation methods and improves on them with the use of auxiliary information and a regression approach. We prove that the new estimator is asymptotically unbiased, and provide a consistent variance estimator to a construct a large sample confidence interval. Finally, we empirically show that the new estimator improves estimation for the current and future policy values, and provides a tight and valid interval estimation in several nonstationary recommendation environments., Comment: AISTATS 2023

Model parallelism is conventionally viewed as a method to scale a single large deep learning model beyond the memory limits of a single device. In this paper, we demonstrate that model parallelism can be additionally used for the statistical multiplexing of multiple devices when serving multiple models, even when a single model can fit into a single device. Our work reveals a fundamental trade-off between the overhead introduced by model parallelism and the opportunity to exploit statistical multiplexing to reduce serving latency in the presence of bursty workloads. We explore the new trade-off space and present a novel serving system, AlpaServe, that determines an efficient strategy for placing and parallelizing collections of large deep learning models across a distributed cluster. Evaluation results on production workloads show that AlpaServe can process requests at up to 10x higher rates or 6x more burstiness while staying within latency constraints for more than 99% of requests., Comment: OSDI 2023

Reachability analysis is a powerful tool when it comes to capturing the behaviour, thus verifying the safety, of autonomous systems. However, general-purpose methods, such as Hamilton-Jacobi approaches, suffer from the curse of dimensionality. In this paper, we mitigate this problem for systems of moderate dimension and we propose a new algorithm based on a tree structure approach with geometric pruning. The numerical examples will include a comparison with a standard finite-difference method for linear and nonlinear problems

The Hofstadter model exemplifies a large class of physical systems characterized by particles hopping on a lattice immersed in a gauge field. Recent advancements on various synthetic platforms have enabled highly-controllable simulations of such systems with tailored gauge fields featuring complex spatial textures. These synthetic gauge fields could introduce synthetic symmetries that do not appear in electronic materials. Here, in an SU(2) non-Abelian Hofstadter model, we theoretically show the emergence of multiple nonsymmorphic chiral symmetries, which combine an internal unitary anti-symmetry with fractional spatial translation. Depending on the values of the gauge fields, the nonsymmorphic chiral symmetries can exhibit non-Abelian algebra and protect Kramer quartet states in the bulk band structure, creating general four-fold degeneracy at all momenta. These nonsymmorphic chiral symmetries protect double Dirac semimetals at zero energy, which become gapped into quantum confined insulating phases upon introducing a boundary. Moreover, the parity of the system size can determine whether the resulting insulating phase is trivial or topological. Our work indicates a pathway for creating topology via synthetic symmetries emergent from synthetic gauge fields.

Compared to fixed-function switches, the flexibility of programmable switches comes at a cost, as programmer mistakes frequently result in subtle bugs in the network data plane. In this paper, we present the design and implementation of FP4, a fuzz-testing framework for P4 switches that achieves high expressiveness, coverage, and scalability. FP4 directly tests running switches by generating semi-random input packets and observing their resulting execution in the data plane. To achieve high coverage and scalability, at runtime, FP4 leverages P4 itself with another "tester" switch that generates and mutates billions of test packets per second entirely in the dataplane. Because testing some program branches requires navigating complex semantic input requirements, FP4 additionally leverages the programmability of P4 by instrumenting the tested program to pass coverage information back to the tester through the packet header. We present case studies showing that FP4 can validate both safety and stateful properties, improves efficiency over existing random packet generation baselines, and reaches 100% coverage in under a minute on a wide range of examples.

Spontaneous symmetry breaking underlies much of our classification of phases of matter and their associated transitions. The nature of the underlying symmetry being broken determines many of the qualitative properties of the phase; this is illustrated by the case of discrete versus continuous symmetry breaking. Indeed, in contrast to the discrete case, the breaking of a continuous symmetry leads to the emergence of gapless Goldstone modes controlling, for instance, the thermodynamic stability of the ordered phase. Here, we realize a two-dimensional dipolar XY model -- which exhibits a continuous spin-rotational symmetry -- utilizing a programmable Rydberg quantum simulator. We demonstrate the adiabatic preparation of correlated low-temperature states of both the XY ferromagnet and the XY antiferromagnet. In the ferromagnetic case, we characterize the presence of long-range XY order, a feature prohibited in the absence of long-range dipolar interaction. Our exploration of the many-body physics of XY interactions complements recent works utilizing the Rydberg-blockade mechanism to realize Ising-type interactions exhibiting discrete spin rotation symmetry., Comment: 18 pages, 3 figures in main text, 9 figures in supplemental methods

Cloud data centers are evolving fast. At the same time, today's large-scale data analytics applications require non-trivial performance tuning that is often specific to the applications, workloads, and data center infrastructure. We propose TeShu, which makes network shuffling an extensible unified service layer common to all data analytics. Since an optimal shuffle depends on a myriad of factors, TeShu introduces parameterized shuffle templates, instantiated by accurate and efficient sampling that enables TeShu to dynamically adapt to different application workloads and data center layouts. Our preliminary experimental results show that TeShu efficiently enables shuffling optimizations that improve performance and adapt to a variety of data center network scenarios., Comment: The technical report of TeShu, which has been accepted at CIDR 2023

The performance of reinforcement learning (RL) agents is sensitive to the choice of hyperparameters. In real-world settings like robotics or industrial control systems, however, testing different hyperparameter configurations directly on the environment can be financially prohibitive, dangerous, or time consuming. We propose a new approach to tune hyperparameters from offline logs of data, to fully specify the hyperparameters for an RL agent that learns online in the real world. The approach is conceptually simple: we first learn a model of the environment from the offline data, which we call a calibration model, and then simulate learning in the calibration model to identify promising hyperparameters. We identify several criteria to make this strategy effective, and develop an approach that satisfies these criteria. We empirically investigate the method in a variety of settings to identify when it is effective and when it fails.

We develop an unsupervised probabilistic model for heterogeneous Electronic Health Record (EHR) data. Utilizing a mixture model formulation, our approach directly models sequences of arbitrary length, such as medications and laboratory results. This allows for subgrouping and incorporation of the dynamics underlying heterogeneous data types. The model consists of a layered set of latent variables that encode underlying structure in the data. These variables represent subject subgroups at the top layer, and unobserved states for sequences in the second layer. We train this model on episodic data from subjects receiving medical care in the Kaiser Permanente Northern California integrated healthcare delivery system. The resulting properties of the trained model generate novel insight from these complex and multifaceted data. In addition, we show how the model can be used to analyze sequences that contribute to assessment of mortality likelihood.

In this paper we investigate the properties of representations learned by deep reinforcement learning systems. Much of the early work on representations for reinforcement learning focused on designing fixed-basis architectures to achieve properties thought to be desirable, such as orthogonality and sparsity. In contrast, the idea behind deep reinforcement learning methods is that the agent designer should not encode representational properties, but rather that the data stream should determine the properties of the representation -- good representations emerge under appropriate training schemes. In this paper we bring these two perspectives together, empirically investigating the properties of representations that support transfer in reinforcement learning. We introduce and measure six representational properties over more than 25 thousand agent-task settings. We consider Deep Q-learning agents with different auxiliary losses in a pixel-based navigation environment, with source and transfer tasks corresponding to different goal locations. We develop a method to better understand why some representations work better for transfer, through a systematic approach varying task similarity and measuring and correlating representation properties with transfer performance. We demonstrate the generality of the methodology by investigating representations learned by a Rainbow agent that successfully transfer across games modes in Atari 2600.

Self-supervised learning algorithms, including BERT and SimCLR, have enabled significant strides in fields like natural language processing, computer vision, and speech processing. However, these algorithms are domain-specific, meaning that new self-supervised learning algorithms must be developed for each new setting, including myriad healthcare, scientific, and multimodal domains. To catalyze progress toward domain-agnostic methods, we introduce DABS: a Domain-Agnostic Benchmark for Self-supervised learning. To perform well on DABS, an algorithm is evaluated on seven diverse domains: natural images, multichannel sensor data, English text, speech recordings, multilingual text, chest x-rays, and images with text descriptions. Each domain contains an unlabeled dataset for pretraining; the model is then is scored based on its downstream performance on a set of labeled tasks in the domain. We also present e-Mix and ShED: two baseline domain-agnostic algorithms; their relatively modest performance demonstrates that significant progress is needed before self-supervised learning is an out-of-the-box solution for arbitrary domains. Code for benchmark datasets and baseline algorithms is available at https://github.com/alextamkin/dabs., Comment: NeurIPS 2021

Offline reinforcement learning -- learning a policy from a batch of data -- is known to be hard for general MDPs. These results motivate the need to look at specific classes of MDPs where offline reinforcement learning might be feasible. In this work, we explore a restricted class of MDPs to obtain guarantees for offline reinforcement learning. The key property, which we call Action Impact Regularity (AIR), is that actions primarily impact a part of the state (an endogenous component) and have limited impact on the remaining part of the state (an exogenous component). AIR is a strong assumption, but it nonetheless holds in a number of real-world domains including financial markets. We discuss algorithms that exploit the AIR property, and provide a theoretical analysis for an algorithm based on Fitted-Q Iteration. Finally, we demonstrate that the algorithm outperforms existing offline reinforcement learning algorithms across different data collection policies in simulated and real world environments where the regularity holds.

Continuously monitoring a wide variety of performance and fault metrics has become a crucial part of operating large-scale datacenter networks. In this work, we ask whether we can reduce the costs to monitor -- in terms of collection, storage and analysis -- by judiciously controlling how much and which measurements we collect. By positing that we can treat almost all measured signals as sampled time-series, we show that we can use signal processing techniques such as the Nyquist-Shannon theorem to avoid wasteful data collection. We show that large savings appear possible by analyzing tens of popular measurements from a production datacenter network. We also discuss the technical challenges that must be solved when applying these techniques in practice.

We theoretically introduce and study a three-dimensional Hofstadter model with linearly varying non-Abelian gauge potentials along all three dimensions. The model can be interpreted as spin-orbit coupling among a trio of Hofstadter butterfly pairs since each Cartesian surface ($xy$, $yz$, or $zx$) of the model reduces to a two-dimensional non-Abelian Hofstadter problem. By evaluating the commutativity among arbitrary loop operators around all axes, we derive its genuine (necessary and sufficient) non-Abelian condition, namely, at least two out of the three hopping phases should be neither 0 nor $\pi$. Under different choices of gauge fields in either the Abelian or the non-Abelian regime, both weak and strong topological insulating phases are identified in the model.

This paper introduces Beldi, a library and runtime system for writing and composing fault-tolerant and transactional stateful serverless functions. Beldi runs on existing providers and lets developers write complex stateful applications that require fault tolerance and transactional semantics without the need to deal with tasks such as load balancing or maintaining virtual machines. Beldi's contributions include extending the log-based fault-tolerant approach in Olive (OSDI 2016) with new data structures, transaction protocols, function invocations, and garbage collection. They also include adapting the resulting framework to work over a federated environment where each serverless function has sovereignty over its own data. We implement three applications on Beldi, including a movie review service, a travel reservation system, and a social media site. Our evaluation on 1,000 AWS Lambdas shows that Beldi's approach is effective and affordable., Comment: Paper appears at OSDI 2020

Catastrophic interference is common in many network-based learning systems, and many proposals exist for mitigating it. But, before we overcome interference we must understand it better. In this work, we provide a definition of interference for control in reinforcement learning. We systematically evaluate our new measures, by assessing correlation with several measures of learning performance, including stability, sample efficiency, and online and offline control performance across a variety of learning architectures. Our new interference measure allows us to ask novel scientific questions about commonly used deep learning architectures. In particular we show that target network frequency is a dominating factor for interference, and that updates on the last layer result in significantly higher interference than updates internal to the network. This new measure can be expensive to compute; we conclude with motivation for an efficient proxy measure and empirically demonstrate it is correlated with our definition of interference., Comment: 18 pages

Clinical researchers often select among and evaluate risk prediction models using standard machine learning metrics based on confusion matrices. However, if these models are used to allocate interventions to patients, standard metrics calculated from retrospective data are only related to model utility (in terms of reductions in outcomes) under certain assumptions. When predictions are delivered repeatedly throughout time (e.g. in a patient encounter), the relationship between standard metrics and utility is further complicated. Several kinds of evaluations have been used in the literature, but it has not been clear what the target of estimation is in each evaluation. We synthesize these approaches, determine what is being estimated in each of them, and discuss under what assumptions those estimates are valid. We demonstrate our insights using simulated data as well as real data used in the design of an early warning system. Our theoretical and empirical results show that evaluations without interventional data either do not estimate meaningful quantities, require strong assumptions, or are limited to estimating best-case scenario bounds.

Clinical predictive algorithms are increasingly being used to form the basis for optimal treatment policies--that is, to enable interventions to be targeted to the patients who will presumably benefit most. Despite taking advantage of recent advances in supervised machine learning, these algorithms remain, in a sense, blunt instruments--often being developed and deployed without a full accounting of the causal aspects of the prediction problems they are intended to solve. Indeed, in many settings, including among patients at risk of readmission, the riskiest patients may derive less benefit from a preventative intervention compared to those at lower risk. Moreover, targeting an intervention to a population, rather than limiting it to a small group of high-risk patients, may lead to far greater overall utility if the patients with the most modifiable (or preventable) outcomes across the population could be identified. Based on these insights, we introduce a causal machine learning framework that decouples this prediction problem into causal and predictive parts, which clearly delineates the complementary roles of causal inference and prediction in this problem. We estimate treatment effects using causal forests, and characterize treatment effect heterogeneity across levels of predicted risk using these estimates. Furthermore, we show how these effect estimates could be used in concert with the modeled "payoffs" associated with successful prevention of individual readmissions to maximize overall utility. Based on data taken from before and after the implementation of a readmissions prevention intervention at Kaiser Permanente Northern California, our results suggest that nearly four times as many readmissions could be prevented annually with this approach compared to targeting this intervention using predicted risk., Comment: 51 pages; 5 figures and 3 tables