Domain-specific Events Research Articles

Recurrent event query decoder for document-level event extraction

Document-level event extraction is a challenging task in natural language processing, as it involves multiple events within a document and scattered event arguments across sentences. To tackle these challenges, we propose a recurrent event query decoder, i.e., a recurrent module that dynamically updates event queries to capture cross-event dependencies. Our approach then generates arguments by extracting role-argument relations using bilinear mapping, which helps address the issue of scattered arguments. Experimental results demonstrate that our proposed approach outperforms state-of-the-art models on a large-scale public dataset and actual application data, achieving significant improvements in F1-score. In domain-specific event extraction applications, our method achieves higher accuracy with fewer resources compared to general-purpose large language models.

Developing an Undergraduate Career Conference: Leveraging Mentorship to Promote Career Discovery

Are students ready for jobs when it comes time to graduation? This is a common question, and one that is often addressed in the media (e.g., Collie, 2019). Despite psychology being one of the most popular degree plans for undergraduate students (e.g., Higher Education Research Institute, 2008), many students in undergraduate psychology programs fail to see the relevance and value of their degree (Borden and Rajecki, 2000). In this work, we designed, delivered, and assessed a career conference for students in psychology. Intentionally different from a career fair where students seek jobs, this event applied a mentorship-based conference model. In this conference model, in addition to professional development training, industry mentors who work in professional fields related to psychology were invited to provide personal insight on their careers in a small-group format. Critical to this model, students were encouraged and able to ask questions that may not be appropriate for a job fair where hiring is happening. Further, this career model involved intentional connections with our Career Services office, allowing for programmatic delivery of career-based content within the domain-specific event. We provide early empirical evidence that this method of career development supports students in learning about career paths that psychology can lead to, identifying skills that will assist them in finding a career, feeling confident in their ability to network effectively, and feeling more connected with professionals in careers related to psychology. We suggest that this model may be beneficial across disciplines.

How Movie Events Engage Childrens’ Brains to Combine Visual Attention with Domain-Specific Processing Involving Number and Theory of Mind in a Cinematic Arena

Using functional magnetic resonance imaging, we tested the hypothesis that cinematic structure shapes variation in social-cognitive brain activity. Using our film, we completed an exploratory analysis of how activations in the temporal-parietal junction (TPJ), and the intraparietal sulcus (IPS) are shaped by variations in insert shots (e.g., shots showing objects that a character has looked at), and by character entrances and exits. We found that IPS and TPJ consistently responded to insert shots, and the correlation between TPJ and IPS responses significantly predicted the prevalence of belief inferences during the sequence. In addition, TPJ responded significantly to entrances and exits of characters. We also completed a qualitative analysis of moments during a sequence that induced relative peaks in TPJ and IPS responding. These analyses not only demonstrate that consistent brain responses can distinguish among meaningful variations in cinematic events but also that these analyses confirm and refine our understanding of the apparent specializations for visual attention and domain-specific event processing in parietal attention networks.

SportsBuzzer: Detecting Events at Real Time in Twitter using Incremental Clustering

In the recent past, twitter users are highly regarded as social sensors who can report events and Twitter has been widely used to detect social and physical events such as earthquakes and traffic jam . Real time event detection in Twitter is the process of detecting events at real time from live tweet stream as soon as an event has happened. Real time event detection from sports tweets, such as Cricket is an interesting, yet a complex problem. Because, an event detection system needs to collect live sports tweets and should rapidly detect key events such as boundary and catch at real-time when the game is ongoing. In this paper, a novel framework is proposed for detecting key events at real time from live tweets of the Cricket sports domain. Feature vectors of live tweets are created using TF-IDF representation and tweet clusters are discovered using Locality Sensitive Hashing (LSH) where the post rate of each cluster based on the volume of tweets is computed. If the post rate is above the predefined threshold, then a key event recognized from that cluster using our domain specific event lexicon for Cricket sports. The predefined threshold helps to filter out small spikes in the tweets volume. The proposed real-time event detection algorithm is extensively evaluated on 2017 IPL T20 Cricket live tweets using ROC evaluation measure. The experimental results on the performance of the proposed approach show that the LSH approach detects sports events with nearly 90% true positive rate and around 10% false positive rate. The results have also demonstrated the influence of different parameters on the accuracy of the event detection.

Automating the Extraction of Static Content and Dynamic Behaviour from e-Commerce Websites

Abstract: E-commerce website owners rely heavily on analysing and summarising the behaviour of costumers, making efforts to influence user actions and optimize success metrics. Machine learning and data mining techniques have been applied in this field, greatly influencing the Internet marketing activities. When faced with a new e-commerce website, the data scientist starts a process of collecting real-time and historical data about it, analysing and transforming this data in order to get a grasp into the website and its users. Data scientists commonly resort to tracking domain-specific events, requiring code modification of the web pages. This paper proposes an alternative approach to retrieve information from a given e-commerce website, collecting data from the site's structure, retrieving semantic information in predefined locations and analysing user's access logs, thus enabling the development of accurate models for predicting users’ future behaviour. This is accomplished by the application of a web mining process, comprehending the site's structure, content and usage in a pipeline, resulting in a web graph of the website, complemented with a categorization of each page and the website's archetypical user profiles.

Detecting well-being via computerized content analysis of brief diary entries.

Two studies evaluated the correspondence between self-reported well-being and codings of emotion and life content by the Linguistic Inquiry and Word Count (LIWC; Pennebaker, Booth, & Francis, 2011). Open-ended diary responses were collected from 206 participants daily for 3 weeks (Study 1) and from 139 participants twice a week for 8 weeks (Study 2). LIWC negative emotion consistently correlated with self-reported negative emotion. LIWC positive emotion correlated with self-reported positive emotion in Study 1 but not in Study 2. No correlations were observed with global life satisfaction. Using a co-occurrence coding method to combine LIWC emotion codings with life-content codings, we estimated the frequency of positive and negative events in 6 life domains (family, friends, academics, health, leisure, and money). Domain-specific event frequencies predicted self-reported satisfaction in all domains in Study 1 but not consistently in Study 2. We suggest that the correspondence between LIWC codings and self-reported well-being is affected by the number of writing samples collected per day as well as the target period (e.g., past day vs. past week) assessed by the self-report measure. Extensions and possible implications for the analyses of similar types of open-ended data (e.g., social media messages) are discussed. (PsycINFO Database Record (c) 2013 APA, all rights reserved).

IDENTIFYING EVENTS TAKING PLACE IN SECOND LIFE VIRTUAL ENVIRONMENTS

Second Life is one of the most popular multi-purpose online virtual worlds, which supports applications in diversified areas relating to real-life activities. Moreover, it is possible to use Second Life in testing Artificial Intelligence theories by creating intelligent virtual agents. For the successful implementation of many of these applications, it is important to accurately identify events taking place inside Second Life. This involves extracting low-level spatio-temporal data and identifying the embedded high-level domain-specific information. This is an aspect that has not been taken into consideration in the prior research related to Second Life. This paper presents a framework that extracts data from Second Life with high accuracy and high frequency, and identifies the high-level domain-specific events and other contextual information embedded in these low-level data. This is guided by our virtual environment formalism, which defines events and states in a virtual environment. This framework is further enhanced to be connected with multiagent development platforms, thus demonstrating its use in the area of Artificial Intelligence.