Causality Detection Research Articles

Cause and Effect: Can Large Language Models Truly Understand Causality?

With the rise of Large Language Models (LLMs), it has become crucial to understand their capabilities and limitations in deciphering and explaining the complex web of causal relationships that language entails. Current methods use either explicit or implicit causal reasoning, yet there is a strong need for a unified approach combining both to tackle a wide array of causal relationships more effectively. This research proposes a novel architecture called Context-Aware Reasoning Enhancement with Counterfactual Analysis (CARE-CA) to enhance causal reasoning and explainability. The proposed framework incorporates an explicit causal detection module with ConceptNet and counterfactual statements, as well as implicit causal detection through LLMs. Our framework goes one step further with a layer of counterfactual explanations to accentuate LLMs' understanding of causality. The knowledge from ConceptNet enhances the performance of multiple causal reasoning tasks such as causal discovery, causal identification, and counterfactual reasoning. The counterfactual sentences add explicit knowledge of `not caused by' scenarios. By combining these powerful modules, our model aims to provide a deeper understanding of causal relationships, enabling enhanced interpretability. Evaluation of benchmark datasets shows improved performance across all metrics, such as accuracy, precision, recall, and F1 scores. We also present CausalNet, a novel dataset specifically curated to benchmark and enhance the causal reasoning capabilities of LLMs. This dataset is accompanied by code designed to facilitate further research in this domain.

Detecting causality in the presence of Byzantine processes: The case of synchronous systems

Detecting causality or the “happens before” relation between events in a distributed system is a fundamental building block for distributed applications. It was recently proved that this problem cannot be solved in an asynchronous distributed system in the presence of Byzantine processes, irrespective of whether the communication mechanism is via unicasts, multicasts, or broadcasts. In light of this impossibility result, we turn attention to synchronous systems and examine the possibility of solving the causality detection problem in such systems. In this paper, we prove that causality detection between events can be solved in the presence of Byzantine processes in a synchronous distributed system. We prove the result by providing two algorithms. The first algorithm uses the Replicated State Machine (RSM) approach and vector clocks. The second algorithm is round-based and uses matrix clocks. The RSM-based algorithm can also run deterministically in partially synchronous systems.

A Comparative Study of Causality Detection Methods in Root Cause Diagnosis: From Industrial Processes to Brain Networks.

Abstracting causal knowledge from process measurements has become an appealing topic for decades, especially for fault root cause analysis (RCA) based on signals recorded by multiple sensors in a complex system. Although many causality detection methods have been developed and applied in different fields, some research communities may have an idiosyncratic implementation of their preferred methods, with limited accessibility to the wider community. Targeting interested experimental researchers and engineers, this paper provides a comprehensive comparison of data-based causality detection methods in root cause diagnosis across two distinct domains. We provide a possible taxonomy of those methods followed by descriptions of the main motivations of those concepts. Of the two cases we investigated, one is a root cause diagnosis of plant-wide oscillations in an industrial process, while the other is the localization of the epileptogenic focus in a human brain network where the connectivity pattern is transient and even more complex. Considering the differences in various causality detection methods, we designed several sets of experiments so that for each case, a total of 11 methods could be appropriately compared under a unified and reasonable evaluation framework. In each case, these methods were implemented separately and in a standard way to infer causal interactions among multiple variables to thus establish the causal network for RCA. From the cross-domain investigation, several findings are presented along with insights into them, including an interpretative pitfall that warrants caution.

Human causality detection and judgment with unsignaled and signaled delayed outcomes.

Four experiments examined human ratings of causal effectiveness, and ability to detect causal relationships, in a nonverbal paradigm. Participants responded on a concurrent random interval, extinction schedule. In the presence of one stimulus, responses produced an outcome (triangle flash); in the presence of the other stimulus, they did not. Following making a judgment of causal effectiveness, two further stimuli were presented simultaneously with one another, and participants had to select one depending on which of the previous two stimuli were associated with effective responses. In all experiments, immediate outcomes were associated with higher causal ratings and better causal detection than outcomes delayed by 3 s. A signal inserted between response and outcome improved ratings and detection (Experiments 2 and 4), even when it was contiguous with the response but not the outcome (Experiments 2 and 3). Stimuli associated with both components (marking cues) did not impact judgments or detection (Experiment 3). Stimuli signaling the availability of an outcome if a response was made (signaled reinforcement) did not improve causal judgments, but did improve detection of stimuli associated with the outcome (Experiment 4). Responses during the delay interfered with detection of the actual relationship when delays were unsignaled (Experiments 1-4), but not with fully or briefly signaled delays (Experiments 2-4), or with signaled reinforcement (Experiment 4). The results suggest a delay stimulus serves to signal the response has been successful and demark the delay period by serving a discriminative function. These findings mirror those seen in nonhuman conditioning. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

High Resilient Asymmetry and Anomaly Detection Based on Data Causality

In the tunnel construction practice, multiple buildings’ tilt rate data are collected. In this study, data causality is defined to reflect the causal relation between the input and output of the building tilt rate detection data. Upon defining and calculating the data causality, a new high resilient causality detection (HiReCau) method is proposed for abnormal building tilt rate detection. A numerical case and another practical case are studied for validation purposes. The case study results show that the proposed HiReCau method can accurately detect high-causality data and low-causality data among the building tilt rate detection data and produces superior results compared with the direct adoption of a machine learning approach. Furthermore, the resilience of HiReCau is validated by investigations testing varied levels of additional low-causality data in the training dataset. Presently, HiReCau is limited to handling problems with a single output. Furthermore, only the back-propagation neural network (BPNN) is tested as the baseline model and there is also room to further expand the data size. The proposed approach is versatile and able to be adjusted to handle fault diagnosis and safety assessment problems in varied theoretical and engineering backgrounds.

Time variability and periodicities of cross‐regional hydroclimatic causation in the contiguous United States

AbstractIdentifying and understanding various causal relations are fundamental to climate dynamics for improving the predictive capacity of Earth system modeling. In particular, causality in Earth systems has manifest temporal periodicities, like physical climate variabilities. To unravel the characteristic frequency of causality in climate dynamics, we develop a data‐analytic framework based on a combination of causality detection and Hilbert spectral analysis, using a long‐term temperature and precipitation dataset in the contiguous United States. Using the Huang–Hilbert transform, we identify the intrinsic frequencies of cross‐regional causality for precipitation and temperature, ranging from interannual to interdecadal time scales. In addition, we analyze the spectra of the physical climate variabilities, including El Niño‐Southern Oscillation and Pacific Decadal Oscillation. It is found that the intrinsic causal frequencies are positively associated with the physics of the oscillations in the global climate system. The proposed methodology provides fresh insights into the causal connectivity in Earth's hydroclimatic system and its underlying mechanism as regulated by the characteristic low‐frequency variability associated with various climatic dynamics.

CausalFD: causal invariance-based fraud detection against camouflaged preference

CausalFD: causal invariance-based fraud detection against camouflaged preference

Robust inference of causality in high-dimensional dynamical processes from the Information Imbalance of distance ranks

We introduce an approach which allows detecting causal relationships between variables for which the time evolution is available. Causality is assessed by a variational scheme based on the Information Imbalance of distance ranks, a statistical test capable of inferring the relative information content of different distance measures. We test whether the predictability of a putative driven system Y can be improved by incorporating information from a potential driver system X, without explicitly modeling the underlying dynamics and without the need to compute probability densities of the dynamic variables. This framework makes causality detection possible even between high-dimensional systems where only few of the variables are known or measured. Benchmark tests on coupled chaotic dynamical systems demonstrate that our approach outperforms other model-free causality detection methods, successfully handling both unidirectional and bidirectional couplings. We also show that the method can be used to robustly detect causality in human electroencephalography data.

Reinforcement learning-based SDN routing scheme empowered by causality detection and GNN.

In recent years, with the rapid development of network applications and the increasing demand for high-quality network service, quality-of-service (QoS) routing has emerged as a critical network technology. The application of machine learning techniques, particularly reinforcement learning and graph neural network, has garnered significant attention in addressing this problem. However, existing reinforcement learning methods lack research on the causal impact of agent actions on the interactive environment, and graph neural network fail to effectively represent link features, which are pivotal for routing optimization. Therefore, this study quantifies the causal influence between the intelligent agent and the interactive environment based on causal inference techniques, aiming to guide the intelligent agent in improving the efficiency of exploring the action space. Simultaneously, graph neural network is employed to embed node and link features, and a reward function is designed that comprehensively considers network performance metrics and causality relevance. A centralized reinforcement learning method is proposed to effectively achieve QoS-aware routing in Software-Defined Networking (SDN). Finally, experiments are conducted in a network simulation environment, and metrics such as packet loss, delay, and throughput all outperform the baseline.

Construction of time series causal network based on partial rank correlation

Discovering causal relationships from time series data is a significant research area in data mining. To tackle the challenges of identifying and constructing causal networks in non-linear, high-dimensional time series data, we introduce the partial rank correlation coefficient and propose two new structure learning algorithms suitable for time-series causal network modeling. In this paper, we make three main contributions. Firstly, we demonstrate the suitability of the partial rank correlation as a criterion for independence testing. Secondly, by integrating the partial rank correlation with constraint-based causal discovery methods, we propose the TS-PRCS algorithm for temporal causal network discovery. Thirdly, we combine ideas from local learning, constraint-based methods, and score-based search to propose the TS-MCHC algorithm, which combines mixed constraints and hill-climbing search for time-series causal discovery. Finally, we validate the effectiveness of the TS-PRCS and TS-MCHC algorithms through experiments on both simulated and real-world data. Compared to existing methods, our proposed algorithms exhibit robust causal detection capabilities in diverse non-linear environments and demonstrate efficient performance in high-dimensional time series networks. The experimental results affirm the ability of our methods to effectively uncover causal relationships among time series data, presenting promising prospects for practical applications.

Radiation, Air Temperature, and Soil Water Availability Drive Tree Water Deficit Across Temporal Scales in Canada's Western Boreal Forest

AbstractChanges are projected for the boreal biome with complex and variable effects on forest vegetation including drought‐induced tree mortality and forest loss. With soil and atmospheric conditions governing drought intensity, specific drivers of trees water stress can be difficult to disentangle across temporal scales. We used wavelet analysis and causality detection to identify potential environmental controls (evapotranspiration, soil moisture, rainfall, vapor pressure deficit, air temperature and photosynthetically active radiation) on daily tree water deficit and on longer periods of tree dehydration in black spruce and tamarack. Daily tree water deficit was controlled by photosynthetically active radiation, vapor pressure deficit, and air temperature, causing greater stand evapotranspiration. Prolonged periods of tree water deficit (multi‐day) were regulated by photosynthetically active radiation and soil moisture. We provide empirical evidence that continued warming and drying will cause short‐term increases in black spruce and tamarack transpiration, but greater drought stress with reduced soil water availability.

Evaluation of metagenomic and pathogen-targeted next-generation sequencing for diagnosis of meningitis and encephalitis in adults: A multicenter prospective observational cohort study in China

Next-generation sequencing (NGS) might aid in the identification of causal pathogens. However, the optimal approaches applied to cerebrospinal fluid (CSF) for detection are unclear, and studies evaluating the application of different NGS workflows for the diagnosis of intracranial infections are limited. In this multicenter, prospective observational cohort study, we described the diagnostic efficacy of pathogen-targeted NGS (ptNGS) and metagenomic NGS (mNGS) compared to that of composite microbiologic assays, for infectious meningitis/encephalitis (M/E). In total, 152 patients diagnosed with clinically suspected M/E at four tertiary hospitals were enrolled; ptNGS and mNGS were used in parallel for pathogen detection in CSF. Among the 89 patients who were diagnosed with definite infectious M/E, 57 and 39 patients had causal microbial detection via ptNGS and mNGS, respectively. The overall accuracy of ptNGS was 65.1%, with a positive percent agreement (PPA) of 64% and a negative percent agreement (NPA) of 66.7%; and the overall accuracy of mNGS was 47.4%, with a PPA of 43.8% and an NPA of 52.4% after discrepancy analysis. There was a significant difference in the detection efficiency between these two methods both for PPA (sensitivity) and overall accuracy for pathogen detection (P<0.05). NGS tests have provided new information in addition to conventional microbiologic tests. ptNGS seems to have superior performance over mNGS for common causative pathogen detection in CSF for infectious M/E.

Modeling and Analyzing Reaction Systems in Maude

Reaction Systems (RSs) are a successful computational framework for modeling systems inspired by biochemistry. An RS defines a set of rules (reactions) over a finite set of entities (e.g., molecules, proteins, genes, etc.). A computation in this system is performed by rewriting a finite set of entities (a computation state) using all the enabled reactions in the RS, thereby producing a new set of entities (a new computation state). The number of entities in the reactions and in the computation states can be large, making the analysis of RS behavior difficult without a proper automated support. In this paper, we use the Maude language—a programming language based on rewriting logic—to define a formal executable semantics for RSs, which can be used to precisely simulate the system behavior as well as to perform reachability analysis over the system computation space. Then, by enriching the proposed semantics, we formalize a forward slicer algorithm for RSs that allows us to observe the evolution of the system on both the initial input and a fragment of it (the slicing criterion), thus facilitating the detection of forward causality and influence relations due to the absence/presence of some entities in the slicing criterion. The pursued approach is illustrated by a biological reaction system that models a gene regulation network for controlling the process of differentiation of T helper lymphocytes.

Surgical therapy and pathogen detection in endogenous endophthalmitis : Analysis of data from five German eye hospitals

Endogenous endophthalmitis results from hematogenous spread of bacterial or fungal infection in severely diseased patients. Specific systemic and intraocular therapy is required. The basis for this treatment is causal pathogen detection in blood culture or vitreous sample. However, functional results are limited. The current article provides practical hints for surgical therapy and pathogen detection in patients with endogenous endophthalmitis. Aretrospective analysis of anonymous data of 68male and female patients from 2018-2023 from five ophthalmology clinics in Germany was performed. Mean age of affected patients was 71.4years (31-96years). Surgical therapy included pars plana vitrectomy (ppV) and intravitreal injection (IVOM). In 44of 68patients (65%), 1-3 surgeries were performed, 4-6surgeries were required in 14/68 (21%) of patients, and 10or more surgeries were required in 4/68patients (6%). Pathogen detection was possible in 34% of vitreous specimens and in 11% of anterior chamber samples. Mean initial visual acuity was logMAR1.5. After treatment and amean follow-up of 2.5months, mean visual acuity was logMAR1.3. Preanalytical methods for specimen collection like the Freiburg endophthalmitis set to optimize pathogen detection are presented. Severe inflammatory intraocular reactions in endogenous endophthalmitis necessitate acombination of ppV and repeated IVOM. In addition to providing avitreous sample, ppV also serves to remove inflammatory fibrin membranes. Early pars plana vitrectomy with specific antibiotic or antifungal therapy should be sought in addition to the focus search and systemic therapy.

The dependence structure and causality detection in crude oil markets

ABSTRACT It is essential to capture the dependence structure among the world’s dominant crude oil markets for understanding the world’s crude oil pricing mechanism. The dependence structure and interaction between Shanghai International Energy Exchange (INE) market and other major crude oil markets in the world are investigated to determine whether the world crude oil market is efficient. Using the vine copula model based on GARCH, we analyze the dependence structure and interactions among these markets. The highest dependence is found between West Texas Intermediate (WTI) and BRENT crude oil markets, displaying strong symmetric tail dependence. DUBAI market exhibits unidirectional Granger causality to WTI and INE markets. Empirical results suggest that the world crude oil markets are not efficient and are far from One Great Pool. DUBAI and Brent markets emerge as leaders in the current international crude oil pricing system, while INE markets do not significantly influence DUBAI and BRENT markets.

A direct approach of causal detection for agriculture related variables via spatial and temporal non-parametric analysis

A direct approach of causal detection for agriculture related variables via spatial and temporal non-parametric analysis

Johansen’s Cointegration Analysis of Some Factors of Economic Growth and Exports of Products from the Republic of Azerbaijan to Ukraine

In the present paper, a Johansen’s cointegration analysis is carried out considering the volume of exports from Azerbaijan to Ukraine, GDP per capita of Ukraine, the openness of the economy of Ukraine and the economically active population of Azerbaijan for the period 1996-2022, also a comparative analysis of the above indicators is carried out, the characteristics of joint short-term and long-term movements are determined. In the course of research the author used the methodology of modified gravitational modeling, the econometric methodology of time series analysis, including tests for checking stationarity, the extended Dickey-Fuller test, the Granger test for the detection of causality, determining the cointegration dependence using the Johansen’s test, a Vector Error Correction Model (VECM) has been built. First, a basic modified model of gravity has been built, and the statistical adequacy of this model has been checked. A comparative analysis of two regression models was carried out after the inclusion of the trend component in one of them. The dynamic structure of regression residuals was studied and the test for heteroscedasticity and autocorrelation was carried out. It is shown that the most suitable specification for cointegration is the quadratic trend in the initial levels with a linear trend in cointegration relationships, which has led to the emergence of two cointegration vectors. As a result of the completed analysis, two cointegration relationships are obtained. The results of the impulse response functions, decomposition of dispersions and the VECM model in the form of combinations of two cointegrating vectors with the expected signs of the adjustment integers showed that the economic indicators used in the analysis for the specified period maintain cointegration in the long term with the not stable equilibrium joint movements of the factors under study.

Detection of Anticipatory Dynamics between a Pair of Zebrafish.

Anticipatory dynamics (AD) is unusual in that responses from an information receiver can appear ahead of triggers from the source, and direction of information flow (DIF) is needed to establish causality. Although it is believed that anticipatory dynamics is important for animals' survival, natural examples are rare. Time series (trajectories) from a pair of interacting zebrafish are used to look for the existence of AD in natural systems. In order to obtain the DIF between the two trajectories, we have made use of a special experimental design to designate information source. However, we have also used common statistical tools such as Granger causality and transfer entropy to detect DIF. In our experiments, we found that a majority of the fish pairs do not show any anticipatory behaviors and only a few pairs displayed possible AD. Interestingly, for fish in this latter group, they do not display AD all the time. Our findings suggest that the formation of schooling of fish might not need the help of AD, and new tools are needed in the detection of causality in AD system.

Learning Entangled Interactions of Complex Causality via Self-Paced Contrastive Learning

Learning causality from large-scale text corpora is an important task with numerous applications—for example, in finance, biology, medicine, and scientific discovery. Prior studies have focused mainly on simple causality, which only includes one cause-effect pair. However, causality is notoriously difficult to understand and analyze because of multiple cause spans and their entangled interactions. To detect complex causality, we propose a self-paced contrastive learning model, namely N2NCause, to learn entangled interactions between multiple spans. Specifically, N2NCause introduces data enhancement operations to convert implicit expressions into explicit expressions with the most rational causal connectives for the synthesis of positive samples and to invert the directed connection between a cause-effect pair for the synthesis of negative samples. To learn the semantic dependency and causal direction of positive and negative samples, self-paced contrastive learning is proposed to learn the entangled interactions among spans, including the interaction direction and interaction field. We evaluated the performance of N2NCause in three cause-effect detection tasks. The experimental results show that, with the least data annotation efforts, N2NCause demonstrates competitive performance in detecting simple cause-effect relations, and it is superior to existing solutions for the detection of complex causality.

Causalized convergent cross-mapping and its approximate equivalence with directed information in causality analysis.

Convergent cross-mapping (CCM) has attracted increased attention recently due to its capability to detect causality in nonseparable systems under deterministic settings, which may not be covered by the traditional Granger causality. From an information-theoretic perspective, causality is often characterized as the directed information (DI) flowing from one side to the other. As information is essentially nondeterministic, a natural question is: does CCM measure DI flow? Here, we first causalize CCM so that it aligns with the presumption in causality analysis-the future values of one process cannot influence the past of the other, and then establish and validate the approximate equivalence of causalized CCM (cCCM) and DI under Gaussian variables through both theoretical derivations and fMRI-based brain network causality analysis. Our simulation result indicates that, in general, cCCM tends to be more robust than DI in causality detection. The underlying argument is that DI relies heavily on probability estimation, which is sensitive to data size as well as digitization procedures; cCCM, on the other hand, gets around this problem through geometric cross-mapping between the manifolds involved. Overall, our analysis demonstrates that cross-mapping provides an alternative way to evaluate DI and is potentially an effective technique for identifying both linear and nonlinear causal coupling in brain neural networks and other settings, either random or deterministic, or both.