Search

A well-known theorem by Fran\c{c}ois Robert expresses the degenerated character of a synchronous Boolean finite dynamical system, in the case where the associated regulatory graph does not contain any circuit: all states of the system go towards a single fixed point. We present a large family of updating modes of Boolean models with the same particularity., Comment: 8 pages, 2 figures

A person is directly racially discriminated against only if her race caused her worse treatment. This implies that race is an attribute sufficiently separable from other attributes to isolate its causal role. But race is embedded in a nexus of social factors that resist isolated treatment. If race is socially constructed, in what sense can it cause worse treatment? Some propose that the perception of race, rather than race itself, causes worse treatment. Others suggest that since causal models require modularity, i.e. the ability to isolate causal effects, attempts to causally model discrimination are misguided. This paper addresses the problem differently. We introduce a framework for reasoning about discrimination, in which race is a high-level abstraction of lower-level features. In this framework, race can be modeled as itself causing worse treatment. Modularity is ensured by allowing assumptions about social construction to be precisely and explicitly stated, via an alignment between race and its constituents. Such assumptions can then be subjected to normative and empirical challenges, which lead to different views of when discrimination occurs. By distinguishing constitutive and causal relations, the abstraction framework pinpoints disagreements in the current literature on modeling discrimination, while preserving a precise causal account of discrimination.

Motivation: In systems biology, modelling strategies aim to decode how molecular components interact to generate dynamical behaviour. Boolean modelling is more and more used, but the description of the dynamics from two-levels components may be too limited to capture certain dynamical properties. %However, in Boolean models, the description of the dynamics may be too limited to capture certain dynamical properties. Multivalued logical models can overcome this limitation by allowing more than two levels for each component. However, multivaluing a Boolean model is challenging. Results: We present MRBM, a method for efficiently identifying the components of a Boolean model to be multivalued in order to capture specific fixed-point reachabilities in the asynchronous dynamics. To this goal, we defined a new updating scheme locating reachability properties in the most permissive dynamics. MRBM is supported by mathematical demonstrations and illustrated on a toy model and on two models of stem cell differentiation.

Ibeling et al. (2023). axiomatize increasingly expressive languages of causation and probability, and Mosse et al. (2024) show that reasoning (specifically the satisfiability problem) in each causal language is as difficult, from a computational complexity perspective, as reasoning in its merely probabilistic or "correlational" counterpart. Introducing a summation operator to capture common devices that appear in applications -- such as the $do$-calculus of Pearl (2009) for causal inference, which makes ample use of marginalization -- van der Zander et al. (2023) partially extend these earlier complexity results to causal and probabilistic languages with marginalization. We complete this extension, fully characterizing the complexity of probabilistic and causal reasoning with summation, demonstrating that these again remain equally difficult. Surprisingly, allowing free variables for random variable values results in a system that is undecidable, so long as the ranges of these random variables are unrestricted. We finally axiomatize these languages featuring marginalization (or more generally summation), resolving open questions posed by Ibeling et al. (2023).

Foundation models such as GPT-4 are fine-tuned to avoid unsafe or otherwise problematic behavior, such as helping to commit crimes or producing racist text. One approach to fine-tuning, called reinforcement learning from human feedback, learns from humans' expressed preferences over multiple outputs. Another approach is constitutional AI, in which the input from humans is a list of high-level principles. But how do we deal with potentially diverging input from humans? How can we aggregate the input into consistent data about "collective" preferences or otherwise use it to make collective choices about model behavior? In this paper, we argue that the field of social choice is well positioned to address these questions, and we discuss ways forward for this agenda, drawing on discussions in a recent workshop on Social Choice for AI Ethics and Safety held in Berkeley, CA, USA in December 2023., Comment: 15 pages, 4 figures

This paper explores the space of (propositional) probabilistic logical languages, ranging from a purely `qualitative' comparative language to a highly `quantitative' language involving arbitrary polynomials over probability terms. While talk of qualitative vs. quantitative may be suggestive, we identify a robust and meaningful boundary in the space by distinguishing systems that encode (at most) additive reasoning from those that encode additive and multiplicative reasoning. The latter includes not only languages with explicit multiplication but also languages expressing notions of dependence and conditionality. We show that the distinction tracks a divide in computational complexity: additive systems remain complete for $\mathsf{NP}$, while multiplicative systems are robustly complete for $\exists\mathbb{R}$. We also address axiomatic questions, offering several new completeness results as well as a proof of non-finite-axiomatizability for comparative probability. Repercussions of our results for conceptual and empirical questions are addressed, and open problems are discussed.

Non-Volatile Memory (NVM) can deliver higher density and lower cost per bit when compared with DRAM. Its main drawback is that it is slower than DRAM. On the other hand, DRAM has scalability problems due to its cost and energy consumption. NVM will likely coexist with DRAM in computer systems and the biggest challenge is to know which data to allocate on each type of memory. A state-of-the-art approach is AutoNUMA, in the Linux kernel. Prior work is limited to measuring AutoNUMA solely in terms of the application execution time, without understanding AutoNUMA's behavior. In this work we provide a more in-depth characterization of AutoNUMA, for instance, identifying where exactly a set of pages are allocated, while keeping track of promotion and demotion decisions performed by AutoNUMA. Our analysis shows that AutoNUMA's benefits can be modest when running graph processing applications, or graph analytics, because most pages have only one access over the entire execution time and other pages accesses have no temporal locality. We make a case for exploring application characteristics using object-level mappings between DRAM and NVM. Our preliminary experiments show that an object-level memory tiering can better capture the application behavior and reduce the execution time of graph analytics by 21% (avg) and 51% (max), when compared to AutoNUMA, while significantly reducing the number of memory accesses in NVM.

Persistent Memory (PMEM), also known as Non-Volatile Memory (NVM), can deliver higher density and lower cost per bit when compared with DRAM. Its main drawback is that it is typically slower than DRAM. On the other hand, DRAM has scalability problems due to its cost and energy consumption. Soon, PMEM will likely coexist with DRAM in computer systems but the biggest challenge is to know which data to allocate on each type of memory. This paper describes a methodology for identifying and characterizing application objects that have the most influence on the application's performance using Intel Optane DC Persistent Memory. In the first part of our work, we built a tool that automates the profiling and analysis of application objects. In the second part, we build a machine learning model to predict the most critical object within large-scale graph-based applications. Our results show that using isolated features does not bring the same benefit compared to using a carefully chosen set of features. By performing data placement using our predictive model, we can reduce the execution time degradation by 12\% (average) and 30\% (max) when compared to the baseline's approach based on LLC misses indicator.

Advances in deep vision techniques and ubiquity of smart cameras will drive the next generation of video analytics. However, video analytics applications consume vast amounts of energy as both deep learning techniques and cameras are power-hungry. In this paper, we focus on a parking video analytics platform and propose RL-CamSleep, a deep reinforcement learning-based technique, to actuate the cameras to reduce the energy footprint while retaining the system's utility. Our key insight is that many video-analytics applications do not always need to be operational, and we can design policies to activate video analytics only when necessary. Moreover, our work is complementary to existing work that focuses on improving hardware and software efficiency. We evaluate our approach on a city-scale parking dataset having 76 streets spread across the city. Our analysis demonstrates how streets have various parking patterns, highlighting the importance of an adaptive policy. Our approach can learn such an adaptive policy that can reduce the average energy consumption by 76.38% and achieve an average accuracy of more than 98% in performing video analytics.

Many tasks in statistical and causal inference can be construed as problems of \emph{entailment} in a suitable formal language. We ask whether those problems are more difficult, from a computational perspective, for \emph{causal} probabilistic languages than for pure probabilistic (or "associational") languages. Despite several senses in which causal reasoning is indeed more complex -- both expressively and inferentially -- we show that causal entailment (or satisfiability) problems can be systematically and robustly reduced to purely probabilistic problems. Thus there is no jump in computational complexity. Along the way we answer several open problems concerning the complexity of well known probability logics, in particular demonstrating the $\exists\mathbb{R}$-completeness of a polynomial probability calculus, as well as a seemingly much simpler system, the logic of comparative conditional probability.

The ideal of a country's health information system (HIS) is to develop processes that ensure easy collection of relevant data and enable their conversion to useful health indicators, which guide decision making and support health interventions. In many Low- and Middle-Income Countries (LMICs), actively engaged in health reform efforts, the role of HIS is crucial, particularly in terms of quality of data and its ability to inspire trust in decision makers to actively use routine HIS data. Recognizing digital platforms potential to support those efforts, several interventions have been implemented in many LMICs. In turn, while the transition from paper registers to digital platforms carries the promise of improving data quality processes, this promise has been notoriously complex to materialize in practice. The authors draw upon more than 15 years of experience implementing HIS in Mozambique to understand how the potential of digital platforms have been realized with respect to data quality, what are the gaps and required remedial steps., Comment: In proceedings of the 1st Virtual Conference on Implications of Information and Digital Technologies for Development, 2021

Many HPC applications suffer from a bottleneck in the shared caches, instruction execution units, I/O or memory bandwidth, even though the remaining resources may be underutilized. It is hard for developers and runtime systems to ensure that all critical resources are fully exploited by a single application, so an attractive technique for increasing HPC system utilization is to colocate multiple applications on the same server. When applications share critical resources, however, contention on shared resources may lead to reduced application performance. In this paper, we show that server efficiency can be improved by first modeling the expected performance degradation of colocated applications based on measured hardware performance counters, and then exploiting the model to determine an optimized mix of colocated applications. This paper presents a new intelligent resource manager and makes the following contributions: (1) a new machine learning model to predict the performance degradation of colocated applications based on hardware counters and (2) an intelligent scheduling scheme deployed on an existing resource manager to enable application co-scheduling with minimum performance degradation. Our results show that our approach achieves performance improvements of 7% (avg) and 12% (max) compared to the standard policy commonly used by existing job managers., Comment: Submitted to Journal of Parallel and Distributed Computing

Recent methods for learning unsupervised visual representations, dubbed contrastive learning, optimize the noise-contrastive estimation (NCE) bound on mutual information between two views of an image. NCE uses randomly sampled negative examples to normalize the objective. In this paper, we show that choosing difficult negatives, or those more similar to the current instance, can yield stronger representations. To do this, we introduce a family of mutual information estimators that sample negatives conditionally -- in a "ring" around each positive. We prove that these estimators lower-bound mutual information, with higher bias but lower variance than NCE. Experimentally, we find our approach, applied on top of existing models (IR, CMC, and MoCo) improves accuracy by 2-5% points in each case, measured by linear evaluation on four standard image datasets. Moreover, we find continued benefits when transferring features to a variety of new image distributions from the Meta-Dataset collection and to a variety of downstream tasks such as object detection, instance segmentation, and keypoint detection., Comment: 8 pages, 4 pages supplement

In recent years, several unsupervised, "contrastive" learning algorithms in vision have been shown to learn representations that perform remarkably well on transfer tasks. We show that this family of algorithms maximizes a lower bound on the mutual information between two or more "views" of an image where typical views come from a composition of image augmentations. Our bound generalizes the InfoNCE objective to support negative sampling from a restricted region of "difficult" contrasts. We find that the choice of negative samples and views are critical to the success of these algorithms. Reformulating previous learning objectives in terms of mutual information also simplifies and stabilizes them. In practice, our new objectives yield representations that outperform those learned with previous approaches for transfer to classification, bounding box detection, instance segmentation, and keypoint detection. % experiments show that choosing more difficult negative samples results in a stronger representation, outperforming those learned with IR, LA, and CMC in classification, bounding box detection, instance segmentation, and keypoint detection. The mutual information framework provides a unifying comparison of approaches to contrastive learning and uncovers the choices that impact representation learning., Comment: 8 pages content; 15 pages supplement with proofs

In modern computer science education, massive open online courses (MOOCs) log thousands of hours of data about how students solve coding challenges. Being so rich in data, these platforms have garnered the interest of the machine learning community, with many new algorithms attempting to autonomously provide feedback to help future students learn. But what about those first hundred thousand students? In most educational contexts (i.e. classrooms), assignments do not have enough historical data for supervised learning. In this paper, we introduce a human-in-the-loop "rubric sampling" approach to tackle the "zero shot" feedback challenge. We are able to provide autonomous feedback for the first students working on an introductory programming assignment with accuracy that substantially outperforms data-hungry algorithms and approaches human level fidelity. Rubric sampling requires minimal teacher effort, can associate feedback with specific parts of a student's solution and can articulate a student's misconceptions in the language of the instructor. Deep learning inference enables rubric sampling to further improve as more assignment specific student data is acquired. We demonstrate our results on a novel dataset from Code.org, the world's largest programming education platform., Comment: To appear at AAAI 2019; 9 pages

A method has been developed for the analysis of images of sentinel lymph nodes generated by a spectral scanning device. The aim is to classify the nodes, excised during surgery for breast cancer, as normal or metastatic. The data from one node constitute spectra at 86 wavelengths for each pixel of a 20*20 grid. For the analysis, the spectra are reduced to scores on two factors, one derived externally from a linear discriminant analysis using spectra taken manually from known normal and metastatic tissue, and one derived from the node under investigation to capture variability orthogonal to the external factor. Then a three-group mixture model (normal, metastatic, non-nodal background) using multivariate t distributions is fitted to the scores, with external data being used to specify informative prior distributions for the parameters of the three distributions. A Markov random field prior imposes smoothness on the image generated by the model. Finally, the node is classified as metastatic if any one pixel in this smoothed image is classified as metastatic. The model parameters were tuned on a training set of nodes, and then the tuned model was tested on a separate validation set of nodes, achieving satisfactory sensitivity and specificity. The aim in developing the analysis was to allow flexibility in the way each node is modelled whilst still using external information. The Bayesian framework employed is ideal for this., Comment: 31 pages, 7 figures

Powerful laser-plasma processes are explored to generate discharge currents of a few $100\,$kA in coil targets, yielding magnetostatic fields (B-fields) in excess of $0.5\,$kT. The quasi-static currents are provided from hot electron ejection from the laser-irradiated surface. According to our model, describing qualitatively the evolution of the discharge current, the major control parameter is the laser irradiance $I_{\mathrm{las}}\lambda_{\mathrm{las}}^2$. The space-time evolution of the B-fields is experimentally characterized by high-frequency bandwidth B-dot probes and by proton-deflectometry measurements. The magnetic pulses, of ns-scale, are long enough to magnetize secondary targets through resistive diffusion. We applied it in experiments of laser-generated relativistic electron transport into solid dielectric targets, yielding an unprecedented 5-fold enhancement of the energy-density flux at $60 \,\mathrm{\mu m}$ depth, compared to unmagnetized transport conditions. These studies pave the ground for magnetized high-energy density physics investigations, related to laser-generated secondary sources of radiation and/or high-energy particles and their transport, to high-gain fusion energy schemes and to laboratory astrophysics., Comment: 11 pages, 7 figures, invited APS

The aim of this work is the investigation of the statistical properties of local electric fields in an ion-electron two component plasmas for coupled conditions. The stochastic fields at a charged or at a neutral point in plasmas involve both slow and fast fluctuation characteristics. The statistical study of these local fields based on a direct time average is done for the first time. For warm and dense plasma conditions, typically $N_{e}\approx 10^{18}cm^{-3}$, $% T_{e}\approx 1eV$, well controlled molecular dynamics (MD) simulations of neutral hydrogen, protons and electrons have been carried out. Relying on these \textit{ab initio} MD calculations this work focuses on an analysis of the concepts of statistically independent slow and fast local field components, based on the consideration of a time averaged electric field. Large differences are found between the results of these MD simulations and corresponding standard results based on static screened fields. The effects discussed are of importance for physical phenomena connected with stochastic electric field fluctuations, e.g., for spectral line broadening in dense plasmas., Comment: 4 pages, 4 figures, submitted to Phys. Rev. Lett

Using a soft pion theorem based on chiral symmetry and a $\Delta(1232)$ resonance model we propose an estimate for the production cross section of low energy pions in the deeply virtual Compton scattering (DVCS) process. In particular, we express the $e p \to e \gamma \pi N$ processes in terms of generalized parton distributions. We provide estimates of the contamination of the $e p \to e \gamma p$ DVCS observables due to this associated pion production processes when the experimental data are not fully exclusive, for a set of kinematical conditions representative of present or planned experiments at JLab, HERMES and COMPASS., Comment: 50 pages, 22 figures