Search

The Problem-oriented AutoML in Clustering (PoAC) framework introduces a novel, flexible approach to automating clustering tasks by addressing the shortcomings of traditional AutoML solutions. Conventional methods often rely on predefined internal Clustering Validity Indexes (CVIs) and static meta-features, limiting their adaptability and effectiveness across diverse clustering tasks. In contrast, PoAC establishes a dynamic connection between the clustering problem, CVIs, and meta-features, allowing users to customize these components based on the specific context and goals of their task. At its core, PoAC employs a surrogate model trained on a large meta-knowledge base of previous clustering datasets and solutions, enabling it to infer the quality of new clustering pipelines and synthesize optimal solutions for unseen datasets. Unlike many AutoML frameworks that are constrained by fixed evaluation metrics and algorithm sets, PoAC is algorithm-agnostic, adapting seamlessly to different clustering problems without requiring additional data or retraining. Experimental results demonstrate that PoAC not only outperforms state-of-the-art frameworks on a variety of datasets but also excels in specific tasks such as data visualization, and highlight its ability to dynamically adjust pipeline configurations based on dataset complexity.

A Language Model is a term that encompasses various types of models designed to understand and generate human communication. Large Language Models (LLMs) have gained significant attention due to their ability to process text with human-like fluency and coherence, making them valuable for a wide range of data-related tasks fashioned as pipelines. The capabilities of LLMs in natural language understanding and generation, combined with their scalability, versatility, and state-of-the-art performance, enable innovative applications across various AI-related fields, including eXplainable Artificial Intelligence (XAI), Automated Machine Learning (AutoML), and Knowledge Graphs (KG). Furthermore, we believe these models can extract valuable insights and make data-driven decisions at scale, a practice commonly referred to as Big Data Analytics (BDA). In this position paper, we provide some discussions in the direction of unlocking synergies among these technologies, which can lead to more powerful and intelligent AI solutions, driving improvements in data pipelines across a wide range of applications and domains integrating humans, computers, and knowledge.

Understanding the decisions of tree-based ensembles and their relationships is pivotal for machine learning model interpretation. Recent attempts to mitigate the human-in-the-loop interpretation challenge have explored the extraction of the decision structure underlying the model taking advantage of graph simplification and path emphasis. However, while these efforts enhance the visualisation experience, they may either result in a visually complex representation or compromise the interpretability of the original ensemble model. In addressing this challenge, especially in complex scenarios, we introduce the Decision Predicate Graph (DPG) as a model-agnostic tool to provide a global interpretation of the model. DPG is a graph structure that captures the tree-based ensemble model and learned dataset details, preserving the relations among features, logical decisions, and predictions towards emphasising insightful points. Leveraging well-known graph theory concepts, such as the notions of centrality and community, DPG offers additional quantitative insights into the model, complementing visualisation techniques, expanding the problem space descriptions, and offering diverse possibilities for extensions. Empirical experiments demonstrate the potential of DPG in addressing traditional benchmarks and complex classification scenarios.

Machine learning models are routinely integrated into process mining pipelines to carry out tasks like data transformation, noise reduction, anomaly detection, classification, and prediction. Often, the design of such models is based on some ad-hoc assumptions about the corresponding data distributions, which are not necessarily in accordance with the non-parametric distributions typically observed with process data. Moreover, the learning procedure they follow ignores the constraints concurrency imposes to process data. Data encoding is a key element to smooth the mismatch between these assumptions but its potential is poorly exploited. In this paper, we argue that a deeper insight into the issues raised by training machine learning models with process data is crucial to ground a sound integration of process mining and machine learning. Our analysis of such issues is aimed at laying the foundation for a methodology aimed at correctly aligning machine learning with process mining requirements and stimulating the research to elaborate in this direction., Comment: 16 pages

Process simulation is gaining attention for its ability to assess potential performance improvements and risks associated with business process changes. The existing literature presents various techniques, generally grounded in process models discovered from event log data or built upon deep learning algorithms. These techniques have specific strengths and limitations. Traditional data-driven approaches offer increased interpretability, while deep learning-based excel at generalizing changes across large event logs. However, the practical application of deep learning faces challenges related to managing stochasticity and integrating information for what-if analysis. This paper introduces a novel recurrent neural architecture tailored to discover COnditioned process Simulation MOdels (CoSMo) based on user-based constraints or any other nature of a-priori knowledge. This architecture facilitates the simulation of event logs that adhere to specific constraints by incorporating declarative-based rules into the learning phase as an attempt to fill the gap of incorporating information into deep learning models to perform what-if analysis. Experimental validation illustrates CoSMo's efficacy in simulating event logs while adhering to predefined declarative conditions, emphasizing both control-flow and data-flow perspectives.

By investigating nonfungible tokens (NFTs), we provide the first systematic study of retail investor behavior through asset bubbles. Given that NFTs are recorded in public blockchains, we are able to track investor behavior over time, leading to the identification of numerous price run-ups and crashes. Our study reveals that agent-level variables, such as investor sophistication, heterogeneity, and wash trading, in addition to aggregate variables, such as volatility, price acceleration, and turnover, significantly predict bubble formation and price crashes. We find that sophisticated investors consistently outperform others and exhibit characteristics consistent with superior information and skills, supporting the narrative surrounding asset pricing bubbles.

Encoding methods are employed across several process mining tasks, including predictive process monitoring, anomalous case detection, trace clustering, etc. These methods are usually performed as preprocessing steps and are responsible for transforming complex information into a numerical feature space. Most papers choose existing encoding methods arbitrarily or employ a strategy based on a specific expert knowledge domain. Moreover, existing methods are employed by using their default hyperparameters without evaluating other options. This practice can lead to several drawbacks, such as suboptimal performance and unfair comparisons with the state-of-the-art. Therefore, this work aims at providing a comprehensive survey on event log encoding by comparing 27 methods, from different natures, in terms of expressivity, scalability, correlation, and domain agnosticism. To the best of our knowledge, this is the most comprehensive study so far focusing on trace encoding in process mining. It contributes to maturing awareness about the role of trace encoding in process mining pipelines and sheds light on issues, concerns, and future research directions regarding the use of encoding methods to bridge the gap between machine learning models and process mining.

We analyze the market quality of centralized crypto exchanges (CEXs) and decentralized blockchain-based venues (DEXs) using a unique and comprehensive dataset. Focusing on two fundamental aspects, transaction costs and deviations from the no-arbitrage condition, we estimate the causal effect of ``gas fees'' on DEX market quality. We show that these fixed costs impose a significant burden on relatively small trades and cause persistent arbitrage deviations. Conversely, DEXs offer more competitive transaction costs for larger trades, offering a more favorable environment for institutional investors. Furthermore, we provide causal evidence that innovations aimed at enhancing the flexibility of liquidity provision in DEX markets lead to sizeable improvements in market quality.

Trace clustering has been extensively used to preprocess event logs. By grouping similar behavior, these techniques guide the identification of sub-logs, producing more understandable models and conformance analytics. Nevertheless, little attention has been posed to the relationship between event log properties and clustering quality. In this work, we propose an Automatic Machine Learning (AutoML) framework to recommend the most suitable pipeline for trace clustering given an event log, which encompasses the encoding method, clustering algorithm, and its hyperparameters. Our experiments were conducted using a thousand event logs, four encoding techniques, and three clustering methods. Results indicate that our framework sheds light on the trace clustering problem and can assist users in choosing the best pipeline considering their scenario., Comment: 17 pages, 7 figures

Process discovery methods have obtained remarkable achievements in Process Mining, delivering comprehensible process models to enhance management capabilities. However, selecting the suitable method for a specific event log highly relies on human expertise, hindering its broad application. Solutions based on Meta-learning (MtL) have been promising for creating systems with reduced human assistance. This paper presents a MtL solution for recommending process discovery methods that maximize model quality according to complementary dimensions. Thanks to our MtL pipeline, it was possible to recommend a discovery method with 92% of accuracy using light-weight features that describe the event log. Our experimental analysis also provided significant insights on the importance of log features in generating recommendations, paving the way to a deeper understanding of the discovery algorithms., Comment: 16 pages, 6 figures

We examine the formation of vortical "smoke rings" as a result of thermalization of energy lost by a jet. We simulate the formation and evolution of these rings using hydrodynamics and define an observable that allows to probe this phenomenon experimentally. We argue that observation of vorticity associated with jets would be an experimental confirmation of the thermalization of the energy lost by quenched jets, and also a probe of shear viscosity., Comment: 8 pages, 7 figures

Relativistic p+A collisions may produce droplets of quark gluon plasma (QGP) that quickly develop a toroidal vortex structure similar to that of an expanding smoke ring. We present viscous relativistic hydrodynamic calculations of ultra-central p+A collisions and develop an experimental observable to probe the structure, correlating the polarization and momentum of hyperons emitted from the collision. The effect is robust against changes in the definition of vorticity used to calculate the polarization. Experiments at RHIC and LHC may test the existence and strength of the vortex toroids, bringing new evidence to bear on the question of collectivity in the smallest QGP droplets., Comment: 5 pages, 4 figures; For v2, references updated; v3 is as published: emphasized that the effect is the same regardless of the "type" of vorticity used in the calculation

Holographic complexity, in the guise of the Complexity = Volume prescription, comes equipped with a natural correspondence between its rate of growth and the average infall momentum of matter in the bulk. This Momentum/Complexity correspondence can be related to an integrated version of the momentum constraint of general relativity. In this paper we propose a generalization, using the full Codazzi equations as a starting point, which successfully accounts for purely gravitational contributions to infall momentum. The proposed formula is explicitly checked in an exact pp-wave solution of the vacuum Einstein equations.

Machine Learning (ML) algorithms have been increasingly applied to problems from several different areas. Despite their growing popularity, their predictive performance is usually affected by the values assigned to their hyperparameters (HPs). As consequence, researchers and practitioners face the challenge of how to set these values. Many users have limited knowledge about ML algorithms and the effect of their HP values and, therefore, do not take advantage of suitable settings. They usually define the HP values by trial and error, which is very subjective, not guaranteed to find good values and dependent on the user experience. Tuning techniques search for HP values able to maximize the predictive performance of induced models for a given dataset, but have the drawback of a high computational cost. Thus, practitioners use default values suggested by the algorithm developer or by tools implementing the algorithm. Although default values usually result in models with acceptable predictive performance, different implementations of the same algorithm can suggest distinct default values. To maintain a balance between tuning and using default values, we propose a strategy to generate new optimized default values. Our approach is grounded on a small set of optimized values able to obtain predictive performance values better than default settings provided by popular tools. After performing a large experiment and a careful analysis of the results, we concluded that our approach delivers better default values. Besides, it leads to competitive solutions when compared to tuned values, making it easier to use and having a lower cost. We also extracted simple rules to guide practitioners in deciding whether to use our new methodology or a HP tuning approach., Comment: 44 pages, 13 figures

The connection between initial-state geometry and anisotropic flow can be quantified through a well-established mapping between $p_T$-integrated flow harmonics and cumulants of the initial transverse energy distribution. In this paper we successfully extend this mapping to also include $p_T$-differential flow. In doing so, we find that subleading principal components of anisotropic flow can reveal previously unobserved details of the hydrodynamic response, in both the linear and the nonlinear regimes. Most importantly, we show that they provide novel information on the small-scale structures present in the initial stage of relativistic heavy-ion collisions., Comment: 18 pages, 12 figures. Published version

We show that the holographic Complexity = Volume proposal satisfies a very general notion of Momentum/Complexity correspondence (PC), based on the Momentum Constraint of General Relativity. It relates the rate of complexity variation with an appropriate matter momentum flux through spacelike extremal surfaces. This formalizes the intuitive idea that `gravitational clumpling' of matter increases complexity, and the required notion of `infall momentum' is shown to have a Newtonian avatar which expresses this idea. The proposed form of the PC correspondence is found to be exact for any solution of Einstein's equations in 2+1 dimensions, and any spherically symmetric solution in arbitrary dimensions, generalizing all previous calculations using spherical thin shells. Gravitational radiation enters through a correction which does not have a straightforward interpretation as a PC correspondence. Other obstructions to an exact PC duality have a topological origin and arise in the presence of wormholes., Comment: 10 pages, 2 figures. Published version

We point out that negative specific heat at high energies is a characteristic feature of many TT-bar deformations, both in the original d = 2 case and in d = 1 quantum mechanical cousins. This note is a contribution to the memorial volume in honor of P. G. O. Freund., Comment: 26 pages, one figure. Added discussion of non-compact CFTs and minor improvements

Machine Learning (ML) algorithms have been used for assessing soil quality parameters along with non-destructive methodologies. Among spectroscopic analytical methodologies, energy dispersive X-ray fluorescence (EDXRF) is one of the more quick, environmentally friendly and less expensive when compared to conventional methods. However, some challenges in EDXRF spectral data analysis still demand more efficient methods capable of providing accurate outcomes. Using Multi-target Regression (MTR) methods, multiple parameters can be predicted, and also taking advantage of inter-correlated parameters the overall predictive performance can be improved. In this study, we proposed the Multi-target Stacked Generalisation (MTSG), a novel MTR method relying on learning from different regressors arranged in stacking structure for a boosted outcome. We compared MTSG and 5 MTR methods for predicting 10 parameters of soil fertility. Random Forest and Support Vector Machine (with linear and radial kernels) were used as learning algorithms embedded into each MTR method. Results showed the superiority of MTR methods over the Single-target Regression (the traditional ML method), reducing the predictive error for 5 parameters. Particularly, MTSG obtained the lowest error for phosphorus, total organic carbon and cation exchange capacity. When observing the relative performance of Support Vector Machine with a radial kernel, the prediction of base saturation percentage was improved in 19%. Finally, the proposed method was able to reduce the average error from 0.67 (single-target) to 0.64 analysing all targets, representing a global improvement of 4.48%., Comment: 20 pages, 5 figures

We establish a version of the Momentum/Complexity (PC) duality between the rate of operator complexity growth and a radial component of bulk momentum for a test system falling into a black hole. In systems of finite entropy, our map remains valid for arbitrarily late times after scrambling. The asymptotic regime of linear complexity growth is associated to a frozen momentum in the interior of the black hole, measured with respect to a time foliation by extremal codimension-one surfaces which saturate without reaching the singularity. The detailed analysis in this paper uses the Volume-Complexity (VC) prescription and an infalling system consisting of a thin shell of dust, but the final PC duality formula should have a much wider degree of generality., Comment: 27 pages, 6 figures

We study the relation between entropy and Action Complexity (AC) for various examples of cosmological singularities in General Relativity. The complexity is defined with respect to the causal domain of dependence of the singular set, and the entropy is evaluated on the boundary of the same causal domain. We find that, contrary to the situation for black hole singularities, the complexity growth near the singularity is controlled by the dynamics of the entropy $S$, with a characteristic linear relation. This formula is found to apply to singularities with vanishing entropy as well as those with diverging entropy. In obtaining these results it is crucial to take into account the AC expansion counterterm, whose associated length scale must be chosen sufficiently large in order to ensure the expected monotonicity properties of the complexity., Comment: 24 pages, 6 figures. Revised version: minor changes in the conclusions. Published version