University Of California , Riverside Research Articles

Discord-based counterfactual explanations for time series classification

AbstractThe opacity inherent in machine learning models presents a significant hindrance to their widespread incorporation into decision-making processes. To address this challenge and foster trust among stakeholders while ensuring decision fairness, the data mining community has been actively advancing the explainable artificial intelligence paradigm. This paper contributes to the evolving field by focusing on counterfactual generation for time series classification models, a domain where research is relatively scarce. We develop, a post-hoc, model agnostic counterfactual explanation algorithm that leverages the Matrix Profile to map time series discords to their nearest neighbors in a target sequence and use this mapping to generate new counterfactual instances. To our knowledge, this is the first effort towards the use of time series discords for counterfactual explanations. We evaluate our algorithm on the University of California Riverside and University of East Anglia archives and compare it to three state-of-the-art univariate and multivariate methods.

Silicone wristbands reveal ubiquitous human exposure to ortho-phthalates and non-ortho-phthalate plasticizers in Southern California

In the United States and abroad, ortho-phthalates and non-ortho-phthalate plasticizers continue to be used within a diverse array of consumer products. Prior California-specific biomonitoring programs for ortho-phthalates have focused on rural, agricultural communities and, to our knowledge, these programs have not measured the potential for exposure to non-ortho-phthalate plasticizers. Therefore, the potential for human exposure to ortho-phthalates and non-ortho-phthalate plasticizers have not been adequately addressed in regions of California that have higher population density. Since there are numerous sources of ortho-phthalates and non-ortho-phthalate plasticizers in population-dense, urban regions, the objective of this study was to leverage silicone wristbands to quantify aggregate ortho-phthalate and non-ortho-phthalate plasticizer exposure over a 5-day period across two different cohorts (2019 and 2020) of undergraduate students at the University of California, Riverside (UCR) that commute from all over Southern California. Based on 5 d of aggregate exposure across two different cohorts, total ortho-phthalate plus non-ortho-phthalate plasticizer concentrations ranged, on average, from ∼100,000–1,000,000 ng/g. Based on the distribution of individual ortho-phthalate and non-ortho-phthalate plasticizer concentrations, the concentrations of di-isononyl phthalate (DiNP, a high molecular weight ortho-phthalate), di (2-ethylhexyl) phthalate (DEHP, a high molecular weight ortho-phthalate), and di-2-ethylhexyl terephthalate (DEHT, a non-ortho-phthalate plasticizer) detected within wristbands were higher than the remaining seven ortho-phthalates and non-ortho-phthalate plasticizers measured, accounting for approximately 94–97% of the total mass depending on the cohort. Overall, our findings raise concerns about chronic DiNP, DEHP, and DEHT exposure in urban, population-dense regions throughout California.

Bake off redux: a review and experimental evaluation of recent time series classification algorithms

AbstractIn 2017, a research paper (Bagnall et al. Data Mining and Knowledge Discovery 31(3):606-660. 2017) compared 18 Time Series Classification (TSC) algorithms on 85 datasets from the University of California, Riverside (UCR) archive. This study, commonly referred to as a ‘bake off’, identified that only nine algorithms performed significantly better than the Dynamic Time Warping (DTW) and Rotation Forest benchmarks that were used. The study categorised each algorithm by the type of feature they extract from time series data, forming a taxonomy of five main algorithm types. This categorisation of algorithms alongside the provision of code and accessible results for reproducibility has helped fuel an increase in popularity of the TSC field. Over six years have passed since this bake off, the UCR archive has expanded to 112 datasets and there have been a large number of new algorithms proposed. We revisit the bake off, seeing how each of the proposed categories have advanced since the original publication, and evaluate the performance of newer algorithms against the previous best-of-category using an expanded UCR archive. We extend the taxonomy to include three new categories to reflect recent developments. Alongside the originally proposed distance, interval, shapelet, dictionary and hybrid based algorithms, we compare newer convolution and feature based algorithms as well as deep learning approaches. We introduce 30 classification datasets either recently donated to the archive or reformatted to the TSC format, and use these to further evaluate the best performing algorithm from each category. Overall, we find that two recently proposed algorithms, MultiROCKET+Hydra (Dempster et al. 2022) and HIVE-COTEv2 (Middlehurst et al. Mach Learn 110:3211-3243. 2021), perform significantly better than other approaches on both the current and new TSC problems.

Field identification of Huanglongbing (HLB) and its management alternatives

EObjective: The objective of this investigation is to identify huanglongbing (HLB) in the field and know their control altenatives. Design/methodology/approach: Was investigated the introduction of the phytopathogen Candidatus Liberibacter, symptoms and control of HLB. The generally observed symptoms of this disease are irregular yellowing in the leaves and reversed ripening on the fruit. Results: According to various authors, an integrated control of the disease must be carried out, such as pest control, production of seedlings in certified nurseries, a suitable nutrition management and working together with educational and research institutions. Limitations on study/implications: HLB or citrus greening is the most devastating disease of citrus trees which has caused millions of dollars in losses worldwide. However, although there are several research, there is still no established cure. Findings/conclusions: Research such as that carried out by geneticist Hailing Jin at the University of California Riverside can give hope for the cure or control of this disease, however, there are limitations to this product since it is not yet commercialized

Reflections on Sikh Dharma at Fifty

ABSTRACT Our panel on Sikh Dharma at Fifty, organized for the May 2023 Sikh studies conference at the University of California-Riverside, allowed me to update what I have previously written on 3HO/Sikh Dharma in light of recent developments affecting the organization. Here I reflect further upon (1) the internal evolution of Sikh Dharma and the self-identities of its followers, (2) the impact of Sikh Dharma – both the organization and its followers – on other Sikhs and Sikhi in North America, and (3) ways in which Sikh studies scholars have written about 3HO/Sikh Dharma positioning for the wider Sikh Panth.

Overjoyed Homegirl

Overjoyed Homegirl

Undergraduates and Political Science Research: Insights from Research Assistants in a Minority-Serving Institution Lab

ABSTRACTThis study examines undergraduate research experiences at a minority-serving institution (MSI) in a political science laboratory. Students contributed to projects in a collaborative research lab at the University of California Riverside that involves undergraduate and graduate students in projects related to health and politics. Adopting a participatory approach to research, the study’s research participants also are coauthors who co-created the research protocols; collected the data; transcribed, coded, and analyzed the data; and wrote up the findings. Our analysis of 12 in-depth interviews with current and former undergraduate research assistants (RAs) found that their work in the lab challenged their perceptions of what research is and what it means to do research; shaped their path to pursue graduate studies; developed their social and professional skills; and offered an inclusive and humanizing experience with graduate students and faculty members. Challenges that the RAs mentioned included time management, bureaucratic accounting and payroll procedures, and feelings of self-doubt; the lab’s culture of inclusion and independence mitigated some of these challenges. Our findings align with the scholarly literature that suggests collaborative research opportunities can have beneficial outcomes, particularly for students from groups that are underrepresented in doctoral programs.

Self-Attention Causal Dilated Convolutional Neural Network for Multivariate Time Series Classification and Its Application

Time Series Classification (TSC) in data mining is gradually developing as an important research direction. Many researchers have developed an extensive interest in Multivariate Time Series Classification (MTSC). The Self-Attention Causal Dilated Convolutional Neural Network (SACDCNN) is proposed to address the limitations of existing models that perform poorly on classification tasks. It designs the residual and dense blocks based on Causal Dilated Convolution based on the traditional residual and dense networks that still have superior performance after deepening the network hierarchy and the dependence of time series on long-range information. A Self-Attention mechanism (SA) is also incorporated to extract the internal autocorrelation of time series features. Comparison experiments on 20 benchmark University of California, Riverside (UCR) and University of California, Irvine (UCI) datasets with eight high-performance classification models show that the method can improve the classification accuracy of time series datasets. Finally, it was applied to petroleum logging reservoir recognition, and a comparison experiment was conducted on two wells. The results show that SACDCNN is effective and significantly superior. It overcomes the shortcomings of traditional logging interpretation techniques and improves the efficiency and success rate of oil and gas exploration.

A Similarity-based Time Series Source Dataset Selection Method for Transfer Learning

Source datasets are selected with high similarity to target datasets to improve the effect of transfer learning. The low similarity between the source and target datasets may lead to negative transfer. This paper proposes a similarity-based time series source dataset selection method for transfer learning. First, to reduce the complexity of the operation, we use Dynamic time-warped barycentric averaging to obtain the prototype signals of each dataset and convert the dataset similarity calculation into the similarity calculation between the prototype signals. The dynamic time warping (DTW) algorithm, commonly used to calculate time-series signal similarity, does not consider the effect of signal complexity and the amount of data the dataset contains on the signal similarity calculation. This article makes improvements to the above issues. Experiments are carried out on 20-time series datasets archived by the University of California Riverside. Experimental results show that the average transfer effect of source datasets selected by the proposed method is better than the commonly used dataset similarity measurement method based on DTW.

Meta-Feature Fusion for Few-Shot Time Series Classification

Deep learning has been widely adopted for end-to-end time-series classification (TSC). However, the effectiveness of deep learning heavily relies on large-scale data. Thus, deep learning is prone to overfit when only few labeled samples are available. Few-shot learning (FSL) aims to address this issue by learning to generalize to new tasks with few training samples (e.g., one or five samples per class). FSL considers learning good representations crucial to classify accurately using discriminative features. In this study, we propose a framework for few-shot TSC that encodes a time series as different types of images (i.e., Recurrence plot, Markov transition field, and Gramian angular summation/difference field) and train these images to the model using the FSL procedure. Different features of each image enable the model to learn rich information. In addition, we propose temporal-context attention (TCA) and meta-feature fusion (MFF) to maximize the representation ability of these images. TCA incorporates global context of the feature map and highlights pixels having informative relevance with other pixels. After extracting features, MFF refines each feature using different kernels generated based on cross-modality features and fuses the refined features. Finally, the test samples are classified to the nearest class prototype in the embedding space. All experiments are conducted on various N-way K-shot problems. Our framework outperforms state-of-the-art models on 28 standard datasets in the UCR (University of California, Riverside) archive, which is a widely used benchmark dataset in time series classification, from 0.34% up to 29.4%.

Contributors

Contributors

Secondary organic aerosol formation from camphene oxidation: measurements and modeling

Abstract. While camphene is one of the dominant monoterpenes measured in biogenic and pyrogenic emission samples, oxidation of camphene has not been well-studied in environmental chambers and very little is known about its potential to form secondary organic aerosol (SOA). The lack of chamber-derived SOA data for camphene may lead to significant uncertainties in predictions of SOA from oxidation of monoterpenes using existing parameterizations when camphene is a significant contributor to total monoterpenes. Therefore, to advance the understanding of camphene oxidation and SOA formation and to improve representation of camphene in air quality models, a series of experiments was performed in the University of California Riverside environmental chamber to explore camphene SOA mass yields and properties across a range of chemical conditions at atmospherically relevant OH concentrations. The experimental results were compared with modeling simulations obtained using two chemically detailed box models: Statewide Air Pollution Research Center (SAPRC) and Generator for Explicit Chemistry and Kinetics of Organics in the Atmosphere (GECKO-A). SOA parameterizations were derived from the chamber data using both the two-product and volatility basis set (VBS) approaches. Experiments performed with added nitrogen oxides (NOx) resulted in higher SOA mass yields (up to 64 %) than experiments performed without added NOx (up to 28 %). In addition, camphene SOA mass yields increased with SOA mass (Mo) at lower mass loadings, but a threshold was reached at higher mass loadings in which the SOA mass yields no longer increased with Mo. SAPRC modeling of the chamber studies suggested that the higher SOA mass yields at higher initial NOx levels were primarily due to higher production of peroxy radicals (RO2) and the generation of highly oxygenated organic molecules (HOMs) formed through unimolecular RO2 reactions. SAPRC predicted that in the presence of NOx, camphene RO2 reacts with NO and the resultant RO2 undergoes hydrogen (H)-shift isomerization reactions; as has been documented previously, such reactions rapidly add oxygen and lead to products with very low volatility (i.e., HOMs). The end products formed in the presence of NOx have significantly lower volatilities, and higher O : C ratios, than those formed by initial camphene RO2 reacting with hydroperoxyl radicals (HO2) or other RO2. Further analysis reveals the existence of an extreme NOx regime, wherein the SOA mass yield can be suppressed again due to high NO / HO2 ratios. Moreover, particle densities were found to decrease from 1.47 to 1.30 g cm−3 as [HC]0 / [NOx]0 increased and O : C decreased. The observed differences in SOA mass yields were largely explained by the gas-phase RO2 chemistry and the competition between RO2+ HO2, RO2+ NO, RO2+ RO2, and RO2 autoxidation reactions.

A Systematic Approach for Evaluating Artificial Intelligence Models in Industrial Settings

Artificial Intelligence (AI) is one of the hottest topics in our society, especially when it comes to solving data-analysis problems. Industry are conducting their digital shifts, and AI is becoming a cornerstone technology for making decisions out of the huge amount of (sensors-based) data available in the production floor. However, such technology may be disappointing when deployed in real conditions. Despite good theoretical performances and high accuracy when trained and tested in isolation, a Machine-Learning (M-L) model may provide degraded performances in real conditions. One reason may be fragility in treating properly unexpected or perturbed data. The objective of the paper is therefore to study the robustness of seven M-L and Deep-Learning (D-L) algorithms, when classifying univariate time-series under perturbations. A systematic approach is proposed for artificially injecting perturbations in the data and for evaluating the robustness of the models. This approach focuses on two perturbations that are likely to happen during data collection. Our experimental study, conducted on twenty sensors’ datasets from the public University of California Riverside (UCR) repository, shows a great disparity of the models’ robustness under data quality degradation. Those results are used to analyse whether the impact of such robustness can be predictable—thanks to decision trees—which would prevent us from testing all perturbations scenarios. Our study shows that building such a predictor is not straightforward and suggests that such a systematic approach needs to be used for evaluating AI models’ robustness.

Generating Adversarial Samples on Multivariate Time Series using Variational Autoencoders

Classification models for multivariate time series have drawn the interest of many researchers to the field with the objective of developing accurate and efficient models. However, limited research has been conducted on generating adversarial samples for multivariate time series classification models. Adversarial samples could become a security concern in systems with complex sets of sensors. This study proposes extending the existing gradient adversarial transformation network (GATN) in combination with adversarial autoencoders to attack multivariate time series classification models. The proposed model attacks classification models by utilizing a distilled model to imitate the output of the multivariate time series classification model. In addition, the adversarial generator function is replaced with a variational autoencoder to enhance the adversarial samples. The developed methodology is tested on two multivariate time series classification models: 1-nearest neighbor dynamic time warping (1-NN DTW) and a fully convolutional network (FCN). This study utilizes 30 multivariate time series benchmarks provided by the University of East Anglia (UEA) and University of California Riverside (UCR). The use of adversarial autoencoders shows an increase in the fraction of successful adversaries generated on multivariate time series. To the best of our knowledge, this is the first study to explore adversarial attacks on multivariate time series. Additionally, we recommend future research utilizing the generated latent space from the variational autoencoders.

Convolutional Neural Network with an Elastic Matching Mechanism for Time Series Classification

Recently, some researchers adopted the convolutional neural network (CNN) for time series classification (TSC) and have achieved better performance than most hand-crafted methods in the University of California, Riverside (UCR) archive. The secret to the success of the CNN is weight sharing, which is robust to the global translation of the time series. However, global translation invariance is not the only case considered for TSC. Temporal distortion is another common phenomenon besides global translation in time series. The scale and phase changes due to temporal distortion bring significant challenges to TSC, which is out of the scope of conventional CNNs. In this paper, a CNN architecture with an elastic matching mechanism, which is named Elastic Matching CNN (short for EM-CNN), is proposed to address this challenge. Compared with the conventional CNN, EM-CNN allows local time shifting between the time series and convolutional kernels, and a matching matrix is exploited to learn the nonlinear alignment between time series and convolutional kernels of the CNN. Several EM-CNN models are proposed in this paper based on diverse CNN models. The results for 85 UCR datasets demonstrate that the elastic matching mechanism effectively improves CNN performance.

Auto-adaptive multilayer perceptron for univariate time series classification

Time Series Classification (TSC) is an intricate problem that has encountered applications in various science fields. Accordingly, many researchers have presented interesting proposals to tackle the TSC problem. Nevertheless, most methods are hand-crafted to classify specific Time Series (TS) and are computationally expensive even for small data sets. In this paper, we propose a new approach to the Multilayer Perceptron (MLP) for TSC. The main novelty is that the hyperparameters related to batch size and the number of neurons in the hidden layers are auto-adapted according to the TS nature. We carried out an empirical study on 61 benchmark data sets from the University of California, Riverside (UCR). The experimental evaluation revealed that our proposal is competitive when we compare the accuracy versus 14 state-of-the-art methods. A non-parametric statistical test verifies that the proposed MLP ranked in fourth place and can be executed on standard computer equipment, making it simple, accessible, and competitive.

The UCR Minicore: a resource for cowpea research and breeding

AbstractIncorporation of new sources of genetic diversity into plant breeding programs is crucial for continuing to improve yield and quality, as well as tolerance to abiotic and biotic stresses. A minicore (the “University of California, Riverside (UCR) Minicore”) composed of 368 worldwide accessions of cultivated cowpea has been assembled, having been derived from the UCR cowpea collection. High‐density genotyping with 51,128 SNPs followed by principal component and genetic assignment analyses identified six subpopulations in the UCR Minicore, mainly differentiated by cultivar group and geographic origin. All six subpopulations were present to some extent in West African material, suggesting that West Africa is a center of diversity for cultivated cowpea. Additionally, population structure analyses supported two routes of introduction of cowpea into the U.S.: (1) from Spain to the southwest U.S. through Northern Mexico and (2) from Africa to the southeast U.S. via the Caribbean. Genome‐wide association studies (GWAS) narrowed several traits to regions containing strong candidate genes. For example, orthologs of the Arabidopsis FLOWERING LOCUS T lie within a major QTL for flowering time. In summary, this diverse, yet compact cowpea collection constitutes a suitable resource to identify loci controlling complex traits, consequently providing markers to assist with breeding to improve this crop of high relevance to global food and nutritional security.

Bidirectional LSTM-RNN-based hybrid deep learning frameworks for univariate time series classification

Time series classification (TSC) has been around for recent decades as a significant research problem for industry practitioners as well as academic researchers. Due to the rapid increase in temporal data in a wide range of disciplines, an incredible amount of algorithms have been proposed. This paper proposes robust approaches based on state-of-the-art techniques, bidirectional long short-term memory (BiLSTM), fully convolutional network (FCN), and attention mechanism. A BiLSTM considers both forward and backward dependencies, and FCN is proven to be good at feature extraction as a TSC baseline. Therefore, we augment BiLSTM and FCN in a hybrid deep learning architecture, BiLSTM-FCN. Moreover, we similarly explore the use of the attention mechanism to check its efficiency on BiLSTM-FCN and propose another model ABiLSTM-FCN. We validate the performance on 85 datasets from the University of California Riverside (UCR) univariate time series archive. The proposed models are evaluated in terms of classification testing error and f1-score and also provide performance comparison with various existing state-of-the-art techniques. The experimental results show that our proposed models perform comprehensively better than the existing state-of-the-art methods and baselines.

Daniel A. Strickman 1953–2020

The field of Medical Entomology has lost a valued entomologist, contributor, colleague, and friend. I have personally lost a close friend with whom I have spent years with discussing pressing challenges in mosquito biology and control. I first met Dan when we both were graduate students with Professor William R. Horsfall at the University of Illinois at Urbana-Champaign. Dan earned his MS in 1976 and Ph.D. in 1978. We both worked on floodwater mosquitoes. Dan and I used to jokingly argue about who processed more dirt from Allerton Park on the Sangamon River floodplain looking for eggs and oviposition sites. We both learned operational mosquito control as entomologists for the Macon Mosquito Abatement District in Decatur, IL. Dan was a quiet man with a unique sense of humor coupled with an outstanding logic for problem solving. He was naturally inquisitive and definitely not dogmatic. He was open and loved to discuss well thought out mosquito and vector control challenges, especially when based on data, observation, or documented information.Dr. Strickman was interested in insects and arthropods at an early age and began his pursuit towards his future professional career as a biologist/entomologist at Dartmouth College and received his Bachelor of Science degree at the University of California-Riverside in 1974.During his career, Dr. Strickman worked in a broad scope of professional responsibilities including medical entomology, operational pest and vector control, research, and management. His career spanned areas from a field researcher to administrative positions in the US military as Chief of Department of Entomology at Walter Reed Army Institute of Research, and the US Department of Agriculture where he held the position of National Program Leader for Veterinary Medical and Urban Entomology and the Acting Director of Overseas Biological Control Laboratories. Although his dominant research focused on mosquitoes, he also worked and contributed knowledge from his studies of tabanids, chiggers, and rodents. His expertise on the parasites and insect vectors causing rickettsia, malaria, and dengue were also key areas and research challenges that intrigued him. His contributions made a significant impact to our knowledge of these vectors regarding both their biology and control. He authored 115 peer-reviewed articles, 16 book chapters, and four books. During his career, an example of his valued expertise is illustrated in the breadth of positions he held, beginning with the US State Department (Peace Corps), the US military (22 years in the Air Force and Army), mosquito abatement (Santa Clara County, CA), the US Department of Agriculture Agricultural Research Service (National Program Leader for Veterinary, Medical, and Urban Entomology and Director of Overseas Laboratories), and finally the Bill & Melinda Gates Foundation (Senior Program Officer, Malaria and Neglected Tropical Diseases). Even in retirement Dan drove to his entomological roots by volunteering his expertise to the Burke Museum of Natural History at the University of Washington, working to curate and examine the museum's collection of fossil insects.During his career, he spent 10 years in countries outside of the USA, including in Latin America, Southwest Asia, East Asia, and Europe. His awards reflect his quality and contributions, including the US military's Legion of Merit and Bronze Star, the US EPA Bronze Medal, the Presidential GreenGov award, the American Mosquito Control Association's John Belkin Award, and the Society for Vector Ecology's Lifetime Achievement Award. Dan was instrumental in supporting the Society of Vector Ecology's annual student competition as our Gates Foundation representative. Dan served on the Editorial Board of the Journal of the American Mosquito Control Association for 13 years.His most recent book is Mosquitoes of the World, to appear in print in 2020 (D. Strickman, R.C. Wilkerson, and Y.-M. Linton, Johns Hopkins University Press). This final work by Dan will be published posthumously. A tribute for a true medical entomologist, a vector control specialist and biologist. He will be missed by many. We have lost a dedicated scholar, colleague, and friend of the American Mosquito Control Association.

Rootstocks Influence Yield Precocity, Productivity, and Pre-Harvest Fruit Drop of Mandared Pigmented Mandarin

Citrus fruit quality and scion productivity are influenced by the choice of rootstock. We aimed to evaluate the effect of rootstocks on yield and fruit quality of Mandared, a triploid pigmented mandarin. To do so, we established a rootstock field trial on a high pH soil (8.6) in which Mandared was grafted onto 11 rootstocks. These included some standard rootstocks, such as trifoliate orange ((Poncirus trifoliata (L.) Raf.), Troyer citrange (Citrus sinensis (L.) Osb. × P. trifoliata), Swingle citrumelo (Citrus paradisi Macf. × P. trifoliata), and C35 citrange (C. sinensis × P. trifoliata), as well as new releases from the Council for Agricultural Research and Economics (CREA, Acireale, Italy) and the University of California Riverside (UCR). The cumulative yield was measured over five consecutive years, while fruit quality was analyzed for two years. The trees on C35, C57 (Citrus sunki Hort. ex. Tan. × P. trifoliata), and C22 (C. sunki × P. trifoliata), started to set fruits one year earlier than the others. The trees on C57 provided some of the highest cumulative yields and canopy volumes. The production of Mandared grafted onto C57 was double that of Mandared grafted onto Troyer, while Mandared grafted onto C35 and C22 resulted in the best yield efficiency. The trees on Swingle and C57 significantly reduced the pre-harvest fruit drop, to which Mandared is particularly sensitive. However, grafting Mandared onto Swingle resulted in the highest variation among replicates, probably due to its high sensitivity to iron chlorosis. Most of the fruit quality parameters, such as fruit size, total soluble solids (TSS), and acidity were not significantly different among the rootstock treatments. However, fruits produced by Mandared grafted onto C22 had one of the highest rates of anthocyanin accumulation. The results indicate that C57, C35, and C22 were the most suitable rootstocks for Mandared in South-Eastern Sicily.