Neural team recommendation has achieved state-of-the-art performance in forming teams of experts whose success in completing complex tasks is almost surely guaranteed.The proposed models frame the problem as a Boolean multilabel classification, mapping the dense vector representations of required skills to the sparse occurrence (multi-hot) vector representation of an optimum subset of experts using multilayer feedforward neural networks.Such approaches, however, suffer from the curse of sparsity in the highdimensional vector of optimum experts in the output layer.In this paper, we propose to reformulate the team recommendation problem into a sequence prediction task and leverage seq-to-seq models, including transformers, to map an input sequence of the required subset of skills onto an output sequence of the optimum subset of experts.Our experiments on four large-scale datasets from various domains, with distinct distributions of skills in teams, show that the seq-to-seq approach is consistently superior overall in a host of classification and information retrieval metrics.Our codebase is available at https://github.com/fani-lab/OpeNTF/tree/nmt.

Neural team recommendation models have brought state-of-the-art efficacy while enhancing efficiency at recommending collaborative teams of experts who, more likely than not, can solve complex tasks. Yet, they suffer from popularity bias and overfit to a few dominant popular experts and, hence, result in discrimination and reduced visibility for already disadvantaged non-popular experts. Such models are trained on randomly shuffled datasets with the disproportionate distribution of a few popular experts over many teams and a sparse long-tailed distribution of non-popular ones, overlooking the difficulty of recommending hard non-popular vs. easy popular experts. To bridge the gap, we propose three curriculum-based learning strategies to empower neural team recommenders sifting through easy popular and hard non-popular experts and to mitigate popularity bias and improve upon the existing neural models. We propose (1) a parametric curriculum that assigns a learnable parameter to each expert enabling the model to learn an expert's levels of difficulty (or conversely, levels of popularity) during training, (2) a parameter-free (non-parametric) curriculum that presumes the worst-case difficulty for each expert based on the model's loss, and (3) a static curriculum to provide a minimum base for comparison amongst curriculum-based learning strategies and lack thereof. Our experiments on two benchmark datasets with distinct distributions of teams over skills showed that our parameter-free curriculum improved the performance of non-variational models across different domains, outperforming its parametric counterpart, and the static curriculum was the poorest. Moreover, among neural models, variational models obtain little to no gain from our proposed curricula, urging further research on more effective curricula for them. The code to reproduce our experiments is publically available at https://github.com/fani-lab/OpeNTF/tree/cl-wsdm25

Query refinement is to enhance the relevance of search results by modifying users' original queries to refined versions. State-of-the-art query refinement models have been trained on web query logs, which are predisposed to topic drifts. To fill the gap, little work has been proposed to generate benchmark datasets of (query  refined query) pairs through an overwhelming application of unsupervised or supervised modifications to the original query while controlling topic drifts. In this paper, however, we propose leveraging natural language backtranslation, a round-trip translation of a query from a source language via target languages, as a simple yet effective unsupervised approach to scale up generating gold-standard benchmark datasets. Backtranslation can (1) uncover terms that are omitted in a query for being commonly understood in a source language, but may not be known in a target language (e.g., figs  (tamil) அத்திமரங்கள்  the fig trees), (2) augment a query with context-aware synonyms in a target language (e.g., italian nobel prize winners  (farsi) برنده های ایتالیایی جایزه نوبل  italian nobel laureates), and (3) help with the semantic disambiguation of polysemous terms and collocations (e.g., custer's last stand  (malay) pertahan terakhir custer  custer's last defence). Our experiments across 5 query sets with different query lengths and topics and 10 languages from 7 language families using 2 neural machine translators validated the effectiveness of query backtranslation in generating a more extensive gold-standard dataset for query refinement. We open-sourced our research at https://github.com/fani-lab/RePair/tree/nqlb.

We present LADy , a Python-based benchmark toolkit to facilitate extracting aspects of products or services in reviews toward which customers target their opinions and sentiments.While there has been a significant increase in aspect-based sentiment analysis, yet the proposed methods' practical implications in real-world settings remain moot for their closed and irreproducible codebases, inability to accommodate datasets from various domains, and poor evaluation methodologies.LADy is an open-source benchmark toolkit with a standard pipeline and experimental details to fill the gaps.It incorporates a host of canonical models along with benchmark datasets from varying domains, including unsolicited online reviews.Leveraging an object-oriented design, LADy readily extends to new models and training datasets.The first of its kind, LADy also features review augmentation via natural language backtranslation that can be integrated into the training phase of the models to boost efficiency and improve efficacy during inference.LADy's codebase, along with the installation instructions and case studies on five datasets for seven methods with backtranslation augmentation over ten languages, can be obtained under cc-by-nc-sa-4.0license at https://github.com/fani-lab/LADy.

Predicting users’ interests on social networks is gaining attention due to its potential to cater customized information and services to the end users. Although previous works have extensively explored how users’ interests can be modeled on social networks, there has been limited investigation into the prediction of users’ future interests. The objective of our work in this paper is to empirically study the effectiveness of different sets of features based on users’ past social interactions, historical interests and their temporal dynamics to predict their interests over a collection of future-yet-unobserved topics. More specifically, we introduce and formalize the features for interest prediction in four categories: user-based, topical, explicit user-topic engagement, and friends’ influence. We further explore the influence of temporality by augmenting features with information pertaining to users’ historical interests and social connections. We model the task of future interest prediction as a learning-to-rank problem where different features and their related categories are ranked based on their relevance and performance in interest prediction, and investigate the efficiency of different features individually and comparatively for predicting the future interest of users with different activity levels in social networks over on unobserved topics. After conducting experiments on a real-world dataset sourced from Twitter, we have identified several noteworthy findings: (1) relevance feature in the category of past explicit user-topic engagement is the strongest indicator for predicting user’s future interest across all user groups, with an observed 8.57% decrease in NDCG and an 8.95% decrease in MAP when it is removed in the ablation study. (2) the observation of an 8.06% decrease in NDCG and a 7.3% decrease in MAP, when topical features such as popularity, freshness, and coherence are removed in the ablation study, highlights their significance as among the strongest indicators for users’ future interest, particularly for low-active users. (3) although temporal features show a clear positive impact across user groups with varying levels of activity (resulting in a 4.5% decrease in NDCG and a 7.3% decrease in MAP when removed in the ablation study), the temporal topical features do not demonstrate a significant positive effect, and 4) The removal of user-specific characteristics such as influence and personality traits in the ablation study reveals their significant impact in predicting future interest over cold topics, reflected by a 5.49% decrease in NDCG and a 5.72% decrease in MAP. Our findings make significant contributions to the field of future interest prediction, offering valuable insights and practical implications for various applications in social network analysis.

Team formation is concerned with the identification of a group of experts who have a high likelihood of effectively collaborating with each other to satisfy a collection of input skills. Solutions to this task have mainly adopted graph operations and at least have the following limitations: (1) they are computationally demanding, as they require finding shortest paths on large collaboration networks; (2) they use various types of heuristics to reduce the exploration space over the collaboration network to become practically feasible; therefore, their results are not necessarily optimal; and (3) they are not well-suited for collaboration network structures given the sparsity of these networks. Our work proposes a variational Bayesian neural network architecture that learns representations for teams whose members have collaborated with each other in the past. The learned representations allow our proposed approach to mine teams that have a past collaborative history and collectively cover the requested desirable set of skills. Through our experiments, we demonstrate that our approach shows stronger performance compared to a range of strong team formation techniques from both quantitative and qualitative perspectives.

Within the context of review analytics, aspects are the features of products and services at which customers target their opinions and sentiments. Aspect detection helps product owners and service providers identify shortcomings and prioritize customers' needs. Existing methods focus on detecting the surface form of an aspect falling short when aspects are latent in reviews, especially in an informal context like in social posts. In this paper, we propose data augmentation via natural language backtranslation to extract latent occurrences of aspects. We presume that backtranslation (1) can reveal latent aspects because they may not be commonly known in the target language and can be generated through backtranslation; (2) augments context-aware synonymous aspects from a target language to the original language, hence addressing the out-of-vocabulary issue; and (3) helps with the semantic disambiguation of polysemous words and collocations. Through our experiments on well-known aspect detection methods across semeval datasets of restaurant and laptop reviews, we demonstrate that review augmentation via backtranslation yields a steady performance boost in baselines. We further contribute LADy at https://github.com/fani-lab/LADy, a benchmark library to support the reproducibility of our research.

Query refinement is the process of transforming users' queries into newrefined versions without semantic drift to enhance the relevance of search results. Prior query refiners were benchmarked on web query logs followingweak assumptions that users' input queries within a search session are about a single topic and improve gradually, which is not necessarily accurate in practice. In this paper, we contribute RePair, an open-source configurable toolkit to generatelarge-scale gold-standard benchmark datasets whose pairs of (original query, refined versions) arealmost surely guaranteed to be in the same semantic context. RePair takes a dataset of queries and their relevance judgements (e.g., msmarco or aol), a sparse or dense retrieval method (e.g., bm25 or colbert ), and an evaluation metric (e.g., map or mrr), and outputs refined versions of queries, each of which with the relevance improvement guarantees under the retrieval method in terms of the evaluation metric. RePair benefits from text-to-text-transfer-transformer (t5) to generate gold-standard datasets for any input query sets and is designed with extensibility in mind. Out of the box, RePair includes gold-standard datasets for aol and msmarco.passage as well as benchmark results of state-of-the-art supervised query suggestion methods on the generated datasets at https://github.com/fani-lab/RePair.

Forming teams of experts who collectively hold a set of required skills and can successfully cooperate is challenging due to the vast pool of feasible candidates with diverse backgrounds, skills, and personalities. Neural models have been proposed to address scalability while maintaining efficacy by learning the distributions of experts and skills from successful teams in the past in order to recommend future teams. However, such models are prone to overfitting when training data suffers from a long-tailed distribution, i.e., few experts have most of the successful collaborations, and the majority has participated sparingly. In this paper, we present an optimization objective that leverages both successful and virtually unsuccessful teams to overcome the long-tailed distribution problem. We propose three negative sampling heuristics that can be seamlessly employed during the training of neural models. We study the synergistic effects of negative samples on the performance of neural models compared to lack thereof on two large-scale benchmark datasets of computer science publications and movies, respectively. Our experiments show that neural models that take unsuccessful teams (negative samples) into account are more efficient and effective in training and inference, respectively.

We contribute OpeNTF, an open-source python-based benchmark library to support neural team formation research. Team formation falls under social information retrieval (Social IR), where the right group of experts should be retrieved to solve a task, which is intractable due to the vast pool of feasible candidates with diverse skills. Even though neural networks could successfully address efficiency while maintaining efficacy, they lack standard implementation and experimental details, which calls for excessive efforts in repeating or reproducing the results in new domains. OpeNTF provides a standard and reproducible platform for neural team formation. It incorporates a host of canonical neural models along with three large-scale training datasets from varying domains. Leveraging an object-oriented structure, OpeNTF readily accommodates the addition of new neural models and training datasets. The first of its kind in neural team formation, OpeNTF also offers negative sampling heuristics that can be seamlessly integrated during model training to boost efficiency and to improve the effectiveness of inference.

Online user communities exhibit distinct temporal dynamics in response to popular topics or breaking events. Despite abundant community detection libraries, there is yet to be one that provides access to the possible user communities in future time intervals. To bridge this gap, we contribute SEERa, an open-source end-to-end community prediction framework to identify future user communities in a text streaming social network. SEERa incorporates state-of-the-art temporal graph neural networks to model inter-user topical affinities at each time interval via streams of temporal graphs. This all takes place while users' topics of interest and hence their inter-user topical affinities are changing over time. SEERa predicts yet-to-be-seen user communities on the final positions of users' vectors in the latent space. Notably, our framework serves as a one-stop-shop to future user communities for Social Information Retrieval and Social Recommendation systems.

We present PyTFL, a library written in Python for the team formation task. In team formation task, the main objective is to form a team of experts given a set of skills. We demonstrate an efficient and well-structured open-source toolkit that can easily be imported into Python. Our toolkit incorporates state-of-the-art approaches for team formation, e.g., neural-based team formation, and supports team formation sub-tasks such as collaboration graph preparation, model training and validation, systematic evaluation based on qualitative and quantitative team metrics, and efficient team formation and prediction. While there are strong research papers on the team formation problem, PyTFL is the first toolkit to be publicly released for this purpose.

Abstract Social interactions through online social media have become a daily routine of many, and the number of those whose real world (offline) and online lives have become intertwined is continuously growing. As such, the interplay of individuals' online and offline activities has been the subject of numerous research studies, the majority of which explored the impact of people's online actions on their offline activities. The opposite direction of impact—the effect of real‐world activities on online actions—has also received attention but to a lesser degree. To contribute to the latter form of impact, this paper reports on a quasi‐experimental design study that examined the presence of causal relations between real‐world activities of online social media users and their online emotional expressions. To this end, we have collected a large dataset (over 17K users) from Twitter and Foursquare, and systematically aligned user content on the two social media platforms. Users' Foursquare check‐ins provided information about their offline activities, whereas the users' expressions of emotions and moods were derived from their Twitter posts. Since our study was based on a quasi‐experimental design, to minimize the impact of covariates, we applied an innovative model of computing propensity scores. Our main findings can be summarized as follows: (a) users' offline activities do impact their affective expressions, both of emotions and moods, as evidenced in their online shared textual content; (b) the impact depends on the type of offline activity and if the user embarks on or abandons the activity. Our findings can be used to devise a personalized recommendation mechanism to help people better manage their online emotional expressions.

We focus on the composition of teams of experts that collectively cover a set of required skills based on their historical collaboration network and expertise. Prior works are primarily based on the shortest path between experts on the expert collaboration network, and suffer from three major shortcomings: (1) they are computationally expensive due to the complexity of finding paths on large network structures; (2) they use a small portion of the entire historical collaboration network to reduce the search space; hence, may form sub-optimal teams; and, (3) they fall short in sparse networks where the majority of the experts have only participated in a few teams in the past. Instead of forming a large network of experts, we propose to learn relationships among experts and skills through a variational Bayes neural architecture wherein: i) we consider all past team compositions as training instances to predict future teams; ii) we bring scalability for large networks of experts due to the neural architecture; and, iii) we address sparsity by incorporating uncertainty on the neural network's parameters which yields a richer representation and more accurate team composition. We empirically demonstrate how our proposed model outperforms the state-of-the-art approaches in terms of effectiveness and efficiency based on a large DBLP dataset.

In this paper, we implement and publicly share a configurable software workflow and a collection of gold standard datasets for training and evaluating supervised query refinement methods. Existing datasets such as AOL and MS MARCO, which have been extensively used in the literature for this purpose, are based on the weak assumption that users' input queries improve gradually within a search session, i.e., the last query where the user ends her information seeking session is the best reconstructed version of her initial query. In practice, such an assumption is not necessarily accurate for a variety of reasons, e.g., topic drift. The objective of our work is to enable researchers to build gold standard query refinement datasets without having to rely on such weak assumptions. Our software workflow, which generates such gold standard query datasets, takes three inputs: (1) a dataset of queries along with their associated relevance judgements (e.g. TREC topics), (2) an information retrieval method (e.g., BM25), and (3) an evaluation metric (e.g., MAP), and outputs a gold standard dataset. The produced gold standard dataset includes a list of revised queries for each query in the input dataset, each of which effectively improves the performance of the specified retrieval method in terms of the desirable evaluation metric. Since our workflow can be used to generate gold standard datasets for any input query set, in this paper, we have generated and publicly shared gold standard datasets for TREC queries associated with Robust04, Gov2, ClueWeb09, and ClueWeb12. The source code of our software workflow, the generated gold datasets, and benchmark results for three state-of-the-art supervised query refinement methods over these datasets are made publicly available for reproducibility purposes.

he abundance of user generated content on social networks pro-vides the opportunity to build models that are able to accurately and effectively extract, mine and predict users' interests with the hopes of enabling more effective user engagement, better quality delivery of appropriate services and higher user satisfaction. While traditional methods for building user profiles relied on AI-based preference elicitation techniques that could have been considered to be intrusive and undesirable by the users, more recent advances are focused on a non-intrusive yet accurate way of determining users' interests and preferences. In this tutorial, we cover five important aspects related to the effective mining of user interests: (1) we introduce the information sources that are used for extracting user interests, (2) various types of user interest profiles that have been proposed in the literature, (3) techniques that have been adopted or proposed for mining user interests, (4) the scalability and re-source requirements of the state of the art methods, and finally (5)the evaluation methodologies that are adopted in the literature for validating the appropriateness of the mined user interest profiles.We also introduce existing challenges, open research question and exciting opportunities for further work.

The abundance of user generated content on social networks provides the opportunity to build models that are able to accurately and effectively extract, mine and predict users' interests with the hopes of enabling more effective user engagement, better quality delivery of appropriate services and higher user satisfaction. While traditional methods for building user profiles relied on AI-based preference elicitation techniques that could have been considered to be intrusive and undesirable by the users, more recent advances are focused on a non-intrusive yet accurate way of determining users' interests and preferences. In this tutorial, we cover five important aspects related to the effective mining of user interests: (1) we introduce the information sources that are used for extracting user interests, (2) various types of user interest profiles that have been proposed in the literature, (3) techniques that have been adopted or proposed for mining user interests, (4) the scalability and resource requirements of the state of the art methods, and finally (5) the evaluation methodologies that are adopted in the literature for validating the appropriateness of the mined user interest profiles. We also introduce existing challenges, open research question and exciting opportunities for further work.

The accurate prediction of users' future topics of interests on social networks can facilitate content recommendation and platform engagement. However, researchers have found that future interest prediction, especially on social networks such as Twitter, is quite challenging due to the rapid changes in community topics and evolution of user interactions. In this context, temporal collaborative filtering methods have already been used to perform user interest prediction, which benefit from similar user behavioral patterns over time to predict how a user's interests might evolve in the future. In this paper, we propose that instead of considering the whole user base within a collaborative filtering framework to predict user interests, it is possible to much more accurately predict such interests by only considering the behavioral patterns of the most influential users related to the user of interest. We model influence as a form of causal dependency between users. To this end, we employ the concept of Granger causality to identify causal dependencies. We show through extensive experimentation that the consideration of only one causally dependent user leads to much more accurate prediction of users' future interests in a host of measures including ranking and rating accuracy metrics.

Researchers have shown that it is possible to identify reported instances of personal life events from users' social content, e.g., tweets. This is known as personal life event detection. In this paper, we take a step forward and explore the possibility of predicting users' next personal life event based solely on the their historically reported personal life events, a task which we refer to as personal life event prediction. We present a framework for modeling streaming social content for the purpose of personal life event prediction and describe how various instantiations of the framework can be developed to build a life event prediction model. In our extensive experiments, we find that (i) historical personal life events of a user have strong predictive power for determining the user's future life event; (ii) the consideration of sequence in historically reported personal life events shows inferior performance compared to models that do not consider sequence, and (iii) the number of historical life events and the length of the past time intervals that are taken into account for making life event predictions can impact prediction performance whereby more recent life events show more relevance for the prediction of future life events.

We propose a neural embedding approach to identify temporally like-minded user communities, i.e., those communities of users who have similar temporal alignment in their topics of interest. Like-minded user communities in social networks are usually identified by either considering explicit structural connections between users (link analysis), users' topics of interest expressed in their posted contents (content analysis), or in tandem. In such communities, however, the users' rich temporal behavior towards topics of interest is overlooked. Only few recent research efforts consider the time dimension and define like-minded user communities as groups of users who share not only similar topical interests but also similar temporal behavior. Temporal like-minded user communities find application in areas such as recommender systems where relevant items are recommended to the users at the right time. In this paper, we tackle the problem of identifying temporally like-minded user communities by leveraging unsupervised feature learning (embeddings). Specifically, we learn a mapping from the user space to a low-dimensional vector space of features that incorporate both topics of interest and their temporal nature. We demonstrate the efficacy of our proposed approach on a Twitter dataset in the context of three applications: news recommendation, user prediction and community selection, where our work is able to outperform the state-of-the-art on important information retrieval metrics.

Researchers have already studied the identification of online communities and the possible impact or influence relationships from several perspectives. For instance, communities of users that are formed based on shared relationships and topological similarities, or communities that consist of users that share similar content. However, little work has been done on detection of communities that simultaneously share topical and temporal similarities. Furthermore, these studies have not explored the causation relationship between the communities. Causation provides systematic explanation as to why communities are formed and helps to predict future communities. This proposal will address two main research questions: i) how can communities that share topical and temporal similarities be identified, and ii) how can causation relation between different online communities be detected and modelled. We model users' behaviour towards topics of interest through multivariate time series to identify like-minded communities. Further, we employ Granger's concept of causality to infer causation between detected communities from corresponding users' time series. Granger causality is the prominent approach in time series modelling and rests on a firm statistical foundation. We assess the proposed community detection methods through comparison with the state of the art and verify the causal model through its prediction accuracy.