Contrastive Learning has recently received interest due to its success in self-supervised representation learning in the computer vision domain. However, the origins of Contrastive Learning date as far back as the 1990s and its development has spanned across many fields and domains including Metric Learning and natural language processing. In this paper, we provide a comprehensive literature review and we propose a general Contrastive Representation Learning framework that simplifies and unifies many different contrastive learning methods. We also provide a taxonomy for each of the components of contrastive learning in order to summarise it and distinguish it from other forms of machine learning. We then discuss the inductive biases which are present in any contrastive learning system and we analyse our framework under different views from various sub-fields of Machine Learning. Examples of how contrastive learning has been applied in computer vision, natural language processing, audio processing, and others, as well as in Reinforcement Learning are also presented. Finally, we discuss the challenges and some of the most promising future research directions ahead.

The Lifelog Search Challenge (LSC) is an annual comparative benchmarking activity for comparing approaches to interactive retrieval from multi-modal lifelogs. LSC'20, the third such challenge, attracts fourteen participants with their interactive lifelog retrieval systems. These systems are comparatively evaluated in front of a live-audience at the LSC workshop at ACM ICMR'20 in Dublin, Ireland. This overview motivates the challenge, presents the dataset and system configuration used in the challenge, and briefly presents the participating teams. CCS CONCEPTS • Human-centered computing → Empirical studies in interaction design; • Information systems → Mobile information processing systems; Search interfaces.

The Implicit Association Test (IAT) is a reaction time based categorization task that measures the differential associative strength between bipolar targets and evaluative attribute concepts as an approach to indexing implicit beliefs or biases. An open question exists as to what exactly the IAT measures, and here EEG (Electroencephalography) has been used to investigate the time course of ERPs (Event-related Potential) indices and implicated brain regions in the IAT. IAT-EEG research identifies a number of early (250–450 ms) negative ERPs indexing early-(pre-response) processing stages of the IAT. ERP activity in this time range is known to index processes related to cognitive control and semantic processing. A central focus of these efforts has been to use IAT-ERPs to delineate the implicit and explicit factors contributing to measured IAT effects. Increasing evidence indicates that cognitive control (and related top-down modulation of attention/perceptual processing) may be components in the effective measurement of IAT effects, as factors such as physical setting or task instruction can change an IAT measurement. In this study we further implicate the role of proactive cognitive control and top-down modulation of attention/perceptual processing in the IAT-EEG. We find statistically significant relationships between D-score (a reaction-time based measure of the IAT-effect) and early ERP-time windows, indicating where more rapid word categorizations driving the IAT effect are present, they are at least partly explainable by neural activity not significantly correlated with the IAT measurement itself. Using LORETA, we identify a number of brain regions driving these ERP-IAT relationships notably involving left-temporal, insular, cingulate, medial frontal and parietal cortex in time regions corresponding to the N2- and P3-related activity. The identified brain regions involved with reduced reaction times on congruent blocks coincide with those of previous studies.

Introduction: There is a growing interest in using generative adversarial networks (GANs) to produce image content that is indistinguishable from real images as judged by a typical person. A number of GAN variants for this purpose have been proposed, however, evaluating GANs performance is inherently difficult because current methods for measuring the quality of their output are not always consistent with what a human perceives. Methods:We propose a novel approach that combines a brain-computer interface (BCI) with GANs to generate a measure we call Neuroscore, which closely mirrors the behavioral ground truth measured from participants tasked with discerning real from synthetic images. This technique we call a neuro-AI interface, as it provides an interface between a human's neural systems and an AI process. In this paper, we first compare the three most widely used metrics in the literature for evaluating GANs in terms of visual quality and compare their outputs with human judgments. Secondly we propose and demonstrate a novel approach using neural signals and rapid serial visual presentation (RSVP) that directly measures a human perceptual response to facial production quality, independent of a behavioral response measurement.Results: The correlation between our proposed Neuroscore and human perceptual judgments has Pearson correlation statistics: r(48) = −0.767, p = 2.089e − 10. We also present the bootstrap result for the correlation i.e., p ≤ 0.0001. Results show that our Neuroscore is more consistent with human judgment compared to the conventional metrics we evaluated.

In almost all countries, precautionary measures are less expensive than medical treatment. The early detection of any disease gives a patient better chances of successful treatment than disease discovery at an advanced stage of its development. If we do not know how to treat patients, any treatment we can provide would be useful and would provide a more comfortable life. Cervical cancer is one such disease, considered to be fourth among the most common types of cancer in women around the world. There are many factors that increase the risk of cervical cancer, such as age and use of hormonal contraceptives. Early detection of cervical cancer helps to raise recovery rates and reduce death rates. This paper aims to use machine learning algorithms to find a model capable of diagnosing cervical cancer with high accuracy and sensitivity. The cervical cancer risk factor dataset from the University of California at Irvine (UCI) was used to construct the classification model through a voting method that combines three classifiers: Decision tree, logistic regression and random forest. The synthetic minority oversampling technique (SMOTE) was used to solve the problem of imbalance dataset and, together with the principal component analysis (PCA) technique, to reduce dimensions that do not affect model accuracy. Then, stratified 10-fold cross-validation technique was used to prevent the overfitting problem. This dataset contains four target variables-Hinselmann, Schiller, Cytology, and Biopsy-with 32 risk factors. We found that using the voting classifier, SMOTE and PCA techniques helped raise the accuracy, sensitivity, and area under the Receiver Operating Characteristic curve (ROC_AUC) of the predictive models created for each of the four target variables to higher rates. In the SMOTE-voting model, accuracy, sensitivity and PPA ratios improved by 0.93 % to 5.13 %, 39.26 % to 46.97 % and 2 % to 29 %, respectively for all target variables. Moreover, using PCA technology reduced computational processing time and increasing model efficiency. Finally, after comparing our results with several previous studies, it was found that our models were able to diagnose cervical cancer more efficiently according to certain evaluation measures.

For the fifth time since 2018, the Lifelog Search Challenge (LSC) facilitated a benchmarking exercise to compare interactive search systems designed for multimodal lifelogs. LSC'22 attracted nine participating research groups who developed interactive lifelog retrieval systems enabling fast and effective access to lifelogs. The systems competed in front of a hybrid audience at the LSC workshop at ACM ICMR'22. This paper presents an introduction to the LSC workshop, the new (larger) dataset used in the competition, and introduces the participating lifelog search systems.

The Lifelog Search Challenge (LSC) is an annual benchmarking challenge for comparing approaches to interactive retrieval from multi-modal lifelogs. LSC'21, the fourth challenge, attracted sixteen participants, each of which had developed interactive retrieval systems for large multimodal lifelogs. These interactive retrieval systems participated in a comparative evaluation in front of an online live-audience at the LSC workshop at ACM ICMR'21. This overview presents the motivation for LSC'21, the lifelog dataset used in the competition, and the participating systems. CCS CONCEPTS• Human-centered computing → Empirical studies in interaction design; • Information systems → Mobile information processing systems; Search interfaces.

No abstract