73 research outputs found
A stigmergy-based analysis of city hotspots to discover trends and anomalies in urban transportation usage
No full textA key aspect of a sustainable urban transportation system is the effectiveness of transportation policies. To be effective, a policy has to consider a broad range of elements, such as pollution emission, traffic flow, and human mobility. Due to the complexity and variability of these elements in the urban area, to produce effective policies remains a very challenging task. With the introduction of the smart city paradigm, a widely available amount of data can be generated in the urban spaces. Such data can be a fundamental source of knowledge to improve policies because they can reflect the sustainability issues underlying the city. In this context, we propose an approach to exploit urban positioning data based on stigmergy, a bio-inspired mechanism providing scalar and temporal aggregation of samples. By employing stigmergy, samples in proximity with each other are aggregated into a functional structure called trail. The trail summarizes relevant dynamics in data and allows matching them, providing a measure of their similarity. Moreover, this mechanism can be specialized to unfold specific dynamics. Specifically, we identify high-density urban areas (i.e. hotspots), analyze their activity over time, and unfold anomalies. Moreover, by matching activity patterns, a continuous measure of the dissimilarity with respect to the typical activity pattern is provided. This measure can be used by policy makers to evaluate the effect of policies and change them dynamically. As a case study, we analyze taxi trip data gathered in Manhattan from 2013 to 2015
Using contrastive learning to inject domain-knowledge into neural networks for recognizing emotions
No full textWith application contexts ranging from psychophysiology to neuromarketing, electroencephalography (EEG)-based
emotion recognition is a fundamental technology for affective
computing. In this context, EEG signals can be processed via
artificial neural networks (NNs) to achieve accurate recognition
of users’ emotions. Still, NNs are rarely employed in realworld decision-making processes, since their internal model
works as a hardly trustable black box. A NN’s reasoning can
be explained in a human-comprehensible manner by exploring
its latent space to understand if some domain knowledge is
actually represented and exploited for the classification. Those
approaches assume that a trained NN autonomously organizes
its latent space according to some domain concepts to process the
data via human-like reasoning. However, there is no guarantee
that such an assumption holds, since the latent space is not
built for this aim. On the other hand, forcing the organization
of the latent space (e.g. via contrastive learning) can result
in poor recognition performances due to information loss. To
guarantee great recognition performances and provide a domainknowledge-driven organization of NNs’ latent space, we combine
the well-known training procedure based on a categorical crossentropy loss with a supervised contrastive learning approach for
continuous values labels. The proposed approach (i) enables the
explanation of NN’s reasoning in terms of the importance of
high-level domain concepts in the final classification, and (ii)
results in a recognition performance comparable to or better than the one achieved via an approach based solely on maximizing
recognition. The proposed approach is tested on the publicly
available MAHNOB datase
Assessing Refugees' Integration via Spatio-temporal Similarities of Mobility and Calling Behaviors
No full textIn Turkey the increasing tension, due to the presence of 3.4 million Syrian refugees, demands the formulation of effective integration policies. Moreover, their design requires tools aimed at understanding the integration of refugees despite the complexity of this phenomenon. In this work, we propose a set of metrics aimed at providing insights and assessing the integration of Syrians refugees, by analyzing a real-world Call Details Records (CDRs) dataset including calls from refugees and locals in Turkey throughout 2017. Specifically, we exploit the similarity between refugees’ and locals’ spatial and temporal behaviors, in terms of communication and mobility in order to assess integration dynamics. Together with the already known methods for data analysis, we use a novel computational approach to analyze spatio-temporal patterns: Computational Stigmergy, a bio-inspired scalar and temporal aggregation of samples. Computational Stigmergy associates each sample to a virtual pheromone deposit (mark). Marks in spatiotemporal proximity are aggregated into functional structures called trails, which summarize the spatiotemporal patterns in data and allows computing the similarity between different patterns. According to our results, collective mobility and behavioral similarity with locals have great potential as measures of integration, since they are: (i) correlated with the amount of interaction with locals; (ii) an effective proxy for refugee's economic capacity, thus refugee's potential employment; and (iii) able to capture events that may disrupt the integration phenomena, such as social tensions
Matching the Expert’s Knowledge via a Counterfactual-Based Feature Importance Measure
No full textTo be employed in real-world applications, explainable artificial intelligence (XAI) techniques need to provide explanations that are comprehensible to experts and decision-makers with no machine learning (ML) background, thus allowing for the validation of the ML model via their domain knowledge.
To this aim, XAI approaches based on feature importance and counterfactuals can be employed, although both have some limitations: the last provide only local explanations, whereas the first can be very computationally expensive. A less computationally-expensive global feature importance measure can be derived by considering the instances close to the model decision boundary and analyzing how often some minor changes in one feature’s values do affect the classification outcome.
However, the validation of XAI techniques in the literature rarely employs the application domain knowledge due to the burden of formalizing it, e.g., providing some degree of expected importance for each feature. Still, given an ML model, it is difficult to determine whether an XAI technique may inject a bias in the explanation (e.g., overestimating or underestimating the importance of a feature) in the absence of such ground truth.
To address this issue, we test our feature importance approach both with the UCI benchmark datasets and real-world smart manufacturing data characterized by annotations provided by domain experts about the expected importance of each feature. If compared to the state-of-the-art, the employed approach results to be reliable and convenient in terms of computation time, as well as more concordant with the expected importance provided by the domain expert
Detecting Permanent and Intermittent Purchase Hotspots via Computational Stigmergy
No full textThe analysis of credit card transactions allows gaining new insights into the spending occurrences and mobility behavior of large numbers of individuals at an unprecedented scale. However, unfolding such spatiotemporal patterns at a community level implies a non-trivial system modeling and parametrization, as well as, a proper representation of the temporal dynamic. In this work we address both those issues by means of a novel computational technique, i.e. computational stigmergy. By using computational stigmergy each sample position is associated with a digital pheromone deposit, which aggregates with other deposits according to their spatiotemporal proximity. By processing transactions data with computational stigmergy, it is possible to identify high-density areas (hotspots) occurring in different time and days, as well as, analyze their consistency over time. Indeed, a hotspot can be permanent, i.e. present throughout the period of observation, or intermittent, i.e. present only in certain time and days due to community level occurrences (e.g. nightlife). Such difference is not only spatial (where the hotspot occurs) and temporal (when the hotspot occurs) but affects also which people visit the hotspot. The proposed approach is tested on a real-world dataset containing the credit card transaction of 60k users between 2014 and 2015
EEG-based motor imagery recognition via novel explainable ensemble learning architecture
No full textBrain–computer interfaces (BCIs) are interactive machines using implicit neurophysiological signals, with applications ranging from medical rehabilitation to smart prostheses and entertainment. In this context, the need for high recognition performance demands increasingly complex machine learning (ML) architectures. Generally, the more complex the architecture, the less transparent its reasoning. This leads to difficulties in motivating their outputs and validating their internal model. Moreover, explainability is explicitly required by recent regulations on personal data processing, which advise against black box modeling. Here, a novel ensemble learning model is proposed aiming to effectively balance recognition performances and explainability. The proposed architecture employs different multilayer perceptrons, each one specialized to distinguish a single pair of classes and to provide counterfactual explanations and the minimal feature changes resulting in a classification shift. Subsequently, their outcomes are weighted to minimize the contribution of the non-competent classifiers and combined to address a multiclass classification problem. Results were gathered from two publicly available datasets on multiclass electroencephalography-based motor imagery and demonstrate that the proposed architecture overcomes state-of-the-art recognition performance while providing information on the most discriminant brain areas and power bands. For the sake of reproducibility, the implementation of the proposed approach is made publicly available
Improving Emotion Recognition Systems by Exploiting the Spatial Information of EEG Sensors
No full textElectroencephalography (EEG)-based emotion recognition is gaining increasing importance due to its potential applications in various scientific fields, ranging from psychophysiology to neuromarketing. A number of approaches have been proposed that use machine learning (ML) technology to achieve high recognition performance, which relies on engineering features from brain activity dynamics. Since ML performance can be improved by utilizing 2D feature representation that exploits the spatial relationships among the features, here we propose a novel input representation that involves re-arranging EEG features as an image that reflects the top view of the subject’s scalp. This approach enables emotion recognition through image-based ML methods such as pre-trained deep neural networks or "trained-from-scratch" convolutional neural networks. We have employed both of these techniques in our study to demonstrate the effectiveness of our proposed input representation. We also compare the recognition performance of these methods against state-of-the-art tabular data analysis approaches, which do not utilize the spatial relationships between the sensors. We test our proposed approach using two publicly available benchmark datasets for EEG-based emotion recognition tasks, namely DEAP and MAHNOB-HCI. Our results show that the "trained-from-scratch" convolutional neural network outperforms the best approaches in the literature, achieving 97.8% and 98.3% accuracy in valence and arousal classification on MAHNOB-HCI, and 91% and 90.4% on DEAP, respectively
From local counterfactuals to global feature importance: efficient, robust, and model-agnostic explanations for brain connectivity networks
No full textBackground: Explainable artificial intelligence (XAI) is a technology that can enhance trust in mental state classifications by providing explanations for the reasoning behind artificial intelligence (AI) models outputs, especially for high-dimensional and highly-correlated brain signals. Feature importance and counterfactual explanations are two common approaches to generate these explanations, but both have drawbacks. While feature importance methods, such as shapley additive explanations (SHAP), can be computationally expensive and sensitive to feature correlation, counterfactual explanations only explain a single outcome instead of the entire model. Methods: To overcome these limitations, we propose a new procedure for computing global feature importance that involves aggregating local counterfactual explanations. This approach is specifically tailored to fMRI signals and is based on the hypothesis that instances close to the decision boundary and their counterfactuals mainly differ in the features identified as most important for the downstream classification task. We refer to this proposed feature importance measure as Boundary Crossing Solo Ratio (BoCSoR), since it quantifies the frequency with which a change in each feature in isolation leads to a change in classification outcome, i.e., the crossing of the model's decision boundary. Results and conclusions: Experimental results on synthetic data and real publicly available fMRI data from the Human Connect project show that the proposed BoCSoR measure is more robust to feature correlation and less computationally expensive than state-of-the-art methods. Additionally, it is equally effective in providing an explanation for the behavior of any AI model for brain signals. These properties are crucial for medical decision support systems, where many different features are often extracted from the same physiological measures and a gold standard is absent. Consequently, computing feature importance may become computationally expensive, and there may be a high probability of mutual correlation among features, leading to unreliable results from state-of-the-art XAI methods
Sleep behavior assessment via smartwatch and stigmergic receptive fields
No full textSleep behavior is a key factor in maintaining good physiological and psychological health. A well-known approach to monitor sleep is polysomnography. However, it is costly and intrusive, which may disturb sleep. Consequently, polysomnography is not suitable for sleep behavior analysis. Other approaches are based on actigraphy and sleep diary. Although being a good source of information for sleep quality assessment, sleep diaries can be affected by cognitive bias related to subject’s sleep perception, while actigraphy overestimates sleep periods and night-time disturbance compared to sleep diaries. Machine learning techniques can improve the objectivity and reliability of the observations. However, since signal morphology vary widely between people, conventional machine learning is complex to set up. In this regard, we present an adaptive, reliable, and innovative computational approach to provide per-night assessment of sleep behavior to the end-user. We exploit heartbeat rate and wrist acceleration data, gathered via smartwatch, in order to identify subject’s sleep behavioral pattern. More specifically, heartbeat rate and wrist motion samples are processed via computational stigmergy, a bio-inspired scalar and temporal aggregation of samples. Stigmergy associates each sample to a digital pheromone deposit (mark) defined in a mono-dimensional space and characterized by evaporation over time. As a consequence, samples close in terms of time and intensity are aggregated into functional structures called trails. The stigmergic trails allow to compute the similarity between time series on different temporal scales, to support classification or clustering processes. The overall computing schema includes a parametric optimization for adapting the structural parameters to individual sleep dynamics. The outcome is a similarity between sleep nights of the same subject, to generate clusters of nights with different quality levels. Experimental results are shown for three real-world subjects. The resulting similarity is also compared with the dynamic time warping, a popular similarity measure for time series
PhysioEx, a new Python library for explainable sleep staging through deep learning
No full textObjective:
Sleep staging is a crucial task in clinical and research contexts for diagnosing and understanding sleep disorders. This work introduces PhysioEx, a Python library designed to support the analysis of sleep stages using deep learning and Explainable AI (XAI).
Approach:
PhysioEx provides an extensible and modular API for standardizing and automating the sleep staging pipeline, covering data preprocessing, model training, testing, fine-tuning, and explainability. It supports both low-resource devices and high-performance computing clusters and includes pretrained models based on the Sleep Heart Health Study (SHHS) dataset. These models support single-channel EEG and multichannel EEG-EOG-EMG configurations and are easily adaptable to custom datasets. PhysioEx also features a command-line interface toolbox allowing users to streamline the model development and deployment. The library offers a range of XAI post-hoc methods to explain model decisions and align them with expert knowledge.
Main results:
PhysioEx benchmark state-of-the-art sleep staging models in a standard pipeline. Enabling a fair comparison between them both on the training source and out-of-domain sources. Its XAI techniques provide insights into deep learning-based sleep staging by linking model decisions to human-understandable concepts, such as AASM-defined rules.
Significance:
PhysioEx addresses the need for a standardized and accessible platform for sleep staging analysis, combining deep learning and XAI. By supporting modular workflows and explainable insights, it bridges the gap between machine learning models and clinical expertise. PhysioEx is publicly available and installable via pip, making it a valuable tool for researchers and practitioners in sleep medicine
- …
