Search

Realistic synthetic tabular data generation encounters significant challenges in preserving privacy, especially when dealing with sensitive information in domains like finance and healthcare. In this paper, we introduce \textit{Federated Tabular Diffusion} (FedTabDiff) for generating high-fidelity mixed-type tabular data without centralized access to the original tabular datasets. Leveraging the strengths of \textit{Denoising Diffusion Probabilistic Models} (DDPMs), our approach addresses the inherent complexities in tabular data, such as mixed attribute types and implicit relationships. More critically, FedTabDiff realizes a decentralized learning scheme that permits multiple entities to collaboratively train a generative model while respecting data privacy and locality. We extend DDPMs into the federated setting for tabular data generation, which includes a synchronous update scheme and weighted averaging for effective model aggregation. Experimental evaluations on real-world financial and medical datasets attest to the framework's capability to produce synthetic data that maintains high fidelity, utility, privacy, and coverage., Comment: 9 pages, 2 figures, 2 tables, preprint version, currently under review

The sharing of microdata, such as fund holdings and derivative instruments, by regulatory institutions presents a unique challenge due to strict data confidentiality and privacy regulations. These challenges often hinder the ability of both academics and practitioners to conduct collaborative research effectively. The emergence of generative models, particularly diffusion models, capable of synthesizing data mimicking the underlying distributions of real-world data presents a compelling solution. This work introduces 'FinDiff', a diffusion model designed to generate real-world financial tabular data for a variety of regulatory downstream tasks, for example economic scenario modeling, stress tests, and fraud detection. The model uses embedding encodings to model mixed modality financial data, comprising both categorical and numeric attributes. The performance of FinDiff in generating synthetic tabular financial data is evaluated against state-of-the-art baseline models using three real-world financial datasets (including two publicly available datasets and one proprietary dataset). Empirical results demonstrate that FinDiff excels in generating synthetic tabular financial data with high fidelity, privacy, and utility., Comment: 9 pages, 5 figures, 3 tables, preprint version, currently under review

The SPARC Toroidal Field Model Coil (TFMC) Program was a three-year effort between 2018 and 2021 that developed novel Rare Earth Yttrium Barium Copper Oxide (REBCO) superconductor technologies and then successfully utilized these technologies to design, build, and test a first-in-class, high-field (~20 T), representative-scale (~3 m) superconducting toroidal field coil. With the principal objective of demonstrating mature, large-scale, REBCO magnets, the project was executed jointly by the MIT Plasma Science and Fusion Center (PSFC) and Commonwealth Fusion Systems (CFS). The TFMC achieved its programmatic goal of experimentally demonstrating a large-scale high-field REBCO magnet, achieving 20.1 T peak field-on-conductor with 40.5 kA of terminal current, 815 kN/m of Lorentz loading on the REBCO stacks, and almost 1 GPa of mechanical stress accommodated by the structural case. Fifteen internal demountable pancake-to-pancake joints operated in the 0.5 to 2.0 nOhm range at 20 K and in magnetic fields up to 12 T. The DC and AC electromagnetic performance of the magnet, predicted by new advances in high-fidelity computational models, was confirmed in two test campaigns while the massively parallel, single-pass, pressure-vessel style coolant scheme capable of large heat removal was validated. The REBCO current lead and feeder system was experimentally qualified up to 50 kA, and the crycooler based cryogenic system provided 600 W of cooling power at 20 K with mass flow rates up to 70 g/s at a maximum design pressure of 20 bar-a for the test campaigns. Finally, the feasibility of using passive, self-protection against a quench in a fusion-scale NI TF coil was experimentally assessed with an intentional open-circuit quench at 31.5 kA terminal current., Comment: 17 pages 9 figures, overview paper and the first of a six-part series of papers covering the TFMC Program

We consider $(1,2)$-rational functions given on the field of $p$-adic numbers $\mathbb Q_p$. In general, such a function has four parameters. We study the case when such a function has two fixed points and show that when there are two fixed points then $(1,2)$-rational function is conjugate to a two-parametric $(1,2)$-rational function. Depending on these two parameters we determine the type of the fixed points, find Siegel disks and the basin of attraction of the fixed points. Moreover, we classify invariant sets and study the ergodicity properties of the function on each invariant set. We describe 2- and 3-periodic orbits of the $p$-adic dynamical systems generated by the two-parametric $(1,2)$-rational functions.

Recent advances in Explainable AI (XAI) increased the demand for deployment of safe and interpretable AI models in various industry sectors. Despite the latest success of deep neural networks in a variety of domains, understanding the decision-making process of such complex models still remains a challenging task for domain experts. Especially in the financial domain, merely pointing to an anomaly composed of often hundreds of mixed type columns, has limited value for experts. Hence, in this paper, we propose a framework for explaining anomalies using denoising autoencoders designed for mixed type tabular data. We specifically focus our technique on anomalies that are erroneous observations. This is achieved by localizing individual sample columns (cells) with potential errors and assigning corresponding confidence scores. In addition, the model provides the expected cell value estimates to fix the errors. We evaluate our approach based on three standard public tabular datasets (Credit Default, Adult, IEEE Fraud) and one proprietary dataset (Holdings). We find that denoising autoencoders applied to this task already outperform other approaches in the cell error detection rates as well as in the expected value rates. Additionally, we analyze how a specialized loss designed for cell error detection can further improve these metrics. Our framework is designed for a domain expert to understand abnormal characteristics of an anomaly, as well as to improve in-house data quality management processes., Comment: 10 pages, 4 figures, 3 tables, preprint version

Detecting accounting anomalies is a recurrent challenge in financial statement audits. Recently, novel methods derived from Deep-Learning (DL) have been proposed to audit the large volumes of a statement's underlying accounting records. However, due to their vast number of parameters, such models exhibit the drawback of being inherently opaque. At the same time, the concealing of a model's inner workings often hinders its real-world application. This observation holds particularly true in financial audits since auditors must reasonably explain and justify their audit decisions. Nowadays, various Explainable AI (XAI) techniques have been proposed to address this challenge, e.g., SHapley Additive exPlanations (SHAP). However, in unsupervised DL as often applied in financial audits, these methods explain the model output at the level of encoded variables. As a result, the explanations of Autoencoder Neural Networks (AENNs) are often hard to comprehend by human auditors. To mitigate this drawback, we propose (RESHAPE), which explains the model output on an aggregated attribute-level. In addition, we introduce an evaluation framework to compare the versatility of XAI methods in auditing. Our experimental results show empirical evidence that RESHAPE results in versatile explanations compared to state-of-the-art baselines. We envision such attribute-level explanations as a necessary next step in the adoption of unsupervised DL techniques in financial auditing., Comment: 9 pages, 4 figures, 5 tables, preprint version, currently under review

The ongoing 'digital transformation' fundamentally changes audit evidence's nature, recording, and volume. Nowadays, the International Standards on Auditing (ISA) requires auditors to examine vast volumes of a financial statement's underlying digital accounting records. As a result, audit firms also 'digitize' their analytical capabilities and invest in Deep Learning (DL), a successful sub-discipline of Machine Learning. The application of DL offers the ability to learn specialized audit models from data of multiple clients, e.g., organizations operating in the same industry or jurisdiction. In general, regulations require auditors to adhere to strict data confidentiality measures. At the same time, recent intriguing discoveries showed that large-scale DL models are vulnerable to leaking sensitive training data information. Today, it often remains unclear how audit firms can apply DL models while complying with data protection regulations. In this work, we propose a Federated Learning framework to train DL models on auditing relevant accounting data of multiple clients. The framework encompasses Differential Privacy and Split Learning capabilities to mitigate data confidentiality risks at model inference. We evaluate our approach to detect accounting anomalies in three real-world datasets of city payments. Our results provide empirical evidence that auditors can benefit from DL models that accumulate knowledge from multiple sources of proprietary client data., Comment: 8 pages, 5 figures, 3 tables, preprint version, currently under review

In this paper the group structure of the $p$-adic ball and sphere are studied. The dynamical system of isometry defined on invariant sphere is investigated. We define the binary operations $\oplus$ and $\odot$ on a ball and sphere respectively, and prove that this sets are compact topological abelian group with respect to the operations. Then we show that any two balls (spheres) with positive radius are isomorphic as groups. We prove that the Haar measure introduced in $\mathbb Z_p$ is also a Haar measure on an arbitrary balls and spheres. We study the dynamical system generated by the isometry defined on a sphere, and show that the trajectory of any initial point that is not a fixed point isn't convergent. We study ergodicity of this $p$-adic dynamical system with respect to normalized Haar measure reduced on the sphere. For $p\geq 3$ we prove that the dynamical systems are not ergodic. But for $p=2$ under some conditions the dynamical system may be ergodic.

Aspiring to achieve an accurate Bitcoin price prediction based on people's opinions on Twitter usually requires millions of tweets, using different text mining techniques (preprocessing, tokenization, stemming, stop word removal), and developing a machine learning model to perform the prediction. These attempts lead to the employment of a significant amount of computer power, central processing unit (CPU) utilization, random-access memory (RAM) usage, and time. To address this issue, in this paper, we consider a classification of tweet attributes that effects on price changes and computer resource usage levels while obtaining an accurate price prediction. To classify tweet attributes having a high effect on price movement, we collect all Bitcoin-related tweets posted in a certain period and divide them into four categories based on the following tweet attributes: $(i)$ the number of followers of the tweet poster, $(ii)$ the number of comments on the tweet, $(iii)$ the number of likes, and $(iv)$ the number of retweets. We separately train and test by using the Q-learning model with the above four categorized sets of tweets and find the best accurate prediction among them. Especially, we design several reward functions to improve the prediction accuracy of the Q-leaning. We compare our approach with a classic approach where all Bitcoin-related tweets are used as input data for the model, by analyzing the CPU workloads, RAM usage, memory, time, and prediction accuracy. The results show that tweets posted by users with the most followers have the most influence on a future price, and their utilization leads to spending 80\% less time, 88.8\% less CPU consumption, and 12.5\% more accurate predictions compared with the classic approach.

In this paper we investigate generalized Gibbs measure (GGM) for $p$-adic Hard-Core(HC) model with a countable set of spin values on a Cayley tree of order $k\geq 2$. This model is defined by $p$-adic parameters $\lambda_i$, $i\in \mathbb N$. We analyze $p$-adic functional equation which provides the consistency condition for the finite-dimensional generalized Gibbs distributions. Each solutions of the functional equation defines a GGM by $p$-adic version of Kolmogorov's theorem. We define $p$-adic Gibbs distributions as limit of the consistent family of finite-dimensional generalized Gibbs distributions and show that, for our $p$-adic HC model on a Cayley tree, such a Gibbs distribution does not exist. Under some conditions on parameters $p$, $k$ and $\lambda_i$ we find the number of translation-invariant and two-periodic GGMs for the $p$-adic HC model on the Cayley tree of order two., Comment: 16 pages

The Nuclotron-base Ion Collider fAcility (NICA) is under construction at the Joint Institute for Nuclear Research (JINR), with commissioning of the facility expected in late 2022. The Multi-Purpose Detector (MPD) has been designed to operate at NICA and its components are currently in production. The detector is expected to be ready for data taking with the first beams from NICA. This document provides an overview of the landscape of the investigation of the QCD phase diagram in the region of maximum baryonic density, where NICA and MPD will be able to provide significant and unique input. It also provides a detailed description of the MPD set-up, including its various subsystems as well as its support and computing infrastructures. Selected performance studies for particular physics measurements at MPD are presented and discussed in the context of existing data and theoretical expectations., Comment: 53 pages, 68 figures, submitted as a Review article to EPJA

We report a fast-track computationally-driven discovery of new SARS-CoV2 Main Protease (M$^{pro}$) inhibitors whose potency range from mM for initial non-covalent ligands to sub-$\mu$M for the final covalent compound (IC50=830 +/- 50 nM). The project extensively relied on high-resolution all-atom molecular dynamics simulations and absolute binding free energy calculations performed using the polarizable AMOEBA force field. The study is complemented by extensive adaptive sampling simulations that are used to rationalize the different ligands binding poses through the explicit reconstruction of the ligand-protein conformation spaces. Machine Learning predictions are also performed to predict selected compound properties. While simulations extensively use High Performance Computing to strongly reduce time-to-solution, they were systematically coupled to Nuclear Magnetic Resonance experiments to drive synthesis and to in vitro characterization of compounds. Such study highlights the power of in silico strategies that rely on structure-based approaches for drug design and allows to address the protein conformational multiplicity problem. The proposed fluorinated tetrahydroquinolines open routes for further optimization of M$^{pro}$ inhibitors towards low nM affinities.

International audit standards require the direct assessment of a financial statement's underlying accounting transactions, referred to as journal entries. Recently, driven by the advances in artificial intelligence, deep learning inspired audit techniques have emerged in the field of auditing vast quantities of journal entry data. Nowadays, the majority of such methods rely on a set of specialized models, each trained for a particular audit task. At the same time, when conducting a financial statement audit, audit teams are confronted with (i) challenging time-budget constraints, (ii) extensive documentation obligations, and (iii) strict model interpretability requirements. As a result, auditors prefer to harness only a single preferably `multi-purpose' model throughout an audit engagement. We propose a contrastive self-supervised learning framework designed to learn audit task invariant accounting data representations to meet this requirement. The framework encompasses deliberate interacting data augmentation policies that utilize the attribute characteristics of journal entry data. We evaluate the framework on two real-world datasets of city payments and transfer the learned representations to three downstream audit tasks: anomaly detection, audit sampling, and audit documentation. Our experimental results provide empirical evidence that the proposed framework offers the ability to increase the efficiency of audits by learning rich and interpretable `multi-task' representations., Comment: 8 pages (excl. appendix), 4 Figures, 3 Tables

This paper is devoted to the complete algebraic and geometric classification of complex $4$-dimensional nilpotent left-symmetric algebras. The corresponding geometric variety has dimension $15$ and decomposes into $3$ irreducible components determined by the Zariski closures of two one-parameter families of algebras and a two-parameter family of algebras (see Theorem B). In particular, there are no rigid $4$-dimensional complex nilpotent left symmetric algebras., Comment: arXiv admin note: substantial text overlap with arXiv:1912.02691, arXiv:1812.01442, arXiv:2004.03598, arXiv:1902.01706

The performance and cost of Bi-2212/Ag wire is limited by the large fraction of silver matrix (~3:1) that is required in the oxide-powder-in-tube fabrication process. An alternative fabrication process is being developed in which fine-powder Bi-2212 is uni-axially compressed to form bars with a thin Ag foil sheath. The fine powder naturally textures (aligns the a-b planes perpendicular to the direction of compaction) with texture >80% using 200 MPa compression. A billet is formed by stacking trapezoidal-cross-section bars in a symmetric 8-12-16 pattern around a Ag rod and enclosing in a Ag-wall extrusion can. The billet is extruded and drawn to fine wire. Results are presented on present status of the development and testing.

REBCO tape has remarkable performance for super-conducting technology, but it is extremely expensive and it is a strongly anisotropic superconductor. The design of a 3.5 T collider dipole is presented in which all turns of the body winding are oriented so that the all turns of REBCO tape operate with maximum current density. Each turn within the winding contains a stack of Cu-clad REBCO tapes; all tapes within the turn are compressed to provide low-resistance Cu-Cu contact, but the successive turns are electrically isolated. Each turn is oriented so that the local magnetic field at the tapes is closely parallel to the tape surface. Sextupole is controlled using a current-programmed turn to provide excellent homogeneity over a 20:1 dynamic range. Issues of current re-distribution during ramping, quench stability, AC losses, synchrotron radiation, and cryogenics have been considered, and appear to provide favorable properties for an ultimate-energy hadron collider.

Blocks-in-Conduit is a novel approach to cable, coil, and splice technologies with unique benefits for a high-current-den-sity toroid D winding to operate at 20 T, 20 K. Blocks of REBCO tape are cabled in conduit, with a thin-wall center-tube spring that provides transposition, stress management at the cable level, and cross-flow cooling of a thick-coil toroid winding. The coil technology utilizes a co-wound armor structure that integrates stress manage-ment and cross-flow cooling and bypasses coil stress to protect the BIC. An interleaved splice joint enables low-resistance demountable splices of the windings. These provisions yield maximum current density in the winding pack, maximum stability of the REBCO tape blocks, and minimum conductor cost for a tokamak toroid.

The audit of financial statements is designed to collect reasonable assurance that an issued statement is free from material misstatement 'true and fair presentation'. International audit standards require the assessment of a statements' underlying accounting relevant transactions referred to as 'journal entries' to detect potential misstatements. To efficiently audit the increasing quantities of such entries, auditors regularly conduct a sample-based assessment referred to as 'audit sampling'. However, the task of audit sampling is often conducted early in the overall audit process. Often at a stage, in which an auditor might be unaware of all generative factors and their dynamics that resulted in the journal entries in-scope of the audit. To overcome this challenge, we propose the application of Vector Quantised-Variational Autoencoder (VQ-VAE) neural networks. We demonstrate, based on two real-world city payment datasets, that such artificial neural networks are capable of learning a quantised representation of accounting data. We show that the learned quantisation uncovers (i) the latent factors of variation and (ii) can be utilised as a highly representative audit sample in financial statement audits., Comment: 8 pages, 5 figures, 3 tables, to appear in Proceedings of the ACM's International Conference on AI in Finance (ICAIF'20), this paper is the initial accepted version