Javier Sanz-Cruzado - Publications

Publications

Here, you shall find the different articles we have published in scientific journals, conferences, workshops, etc. For each category, articles are ordered from more recent to older.

Journal publications

Javier Sanz-Cruzado, Pablo Castells, Craig Macdonald, Iadh Ounis. Effective Contact Recommendation in Social Networks by Adaptation of Information Retrieval Models.
Information Processing & Management , 57 (5), 102285, September 2020. DOI: https://doi.org/10.1016/j.ipm.2020.102285

Abstract: We investigate a novel perspective to the development of effective algorithms for contact recommendation in social networks, where the problem consists of automatically predicting people that a given user may wish or benefit from connecting to in the network. Specifically, we explore the connection between contact recommendation and the text information retrieval (IR) task, by investigating the adaptation of IR models (classical and supervised) for recommending people in social networks, using only the structure of these networks. We first explore the use of adapted unsupervised IR models as direct standalone recommender systems. Seeking additional effectiveness enhancements, we further explore the use of IR models as neighbor selection methods, in place of common similarity measures, in user-based and item-based nearest-neighbors (kNN) collaborative filtering approaches. On top of this, we investigate the application of learning to rank approaches borrowed from text IR to achieve additional improvements. We report thorough experiments over data obtained from Twitter and Facebook where we observe that IR models, particularly BM25, are competitive compared to state-of-the art contact recommendation methods. We provide further empirical analysis of the additional effectiveness that can be achieved by the integration of IR models into kNN and learning to rank schemes. Our research shows that the IR models are effective in three roles: as direct contact recommenders, as neighbor selectors in collaborative filtering and as samplers and features in learning to rank.
Keywords: social networks, recommender systems, contact recommendation, information retrieval models, collaborative filtering, k nearest neighbors, learning to rank.
Code: https://github.com/ir-uam/IR-models-4-contact-recommendation

Conference publications

Youngbin Lee, Yeyin Kim, Javier Sanz-Cruzado, Richard McCreadie, Yongjae Lee. Stock Recommendations for Individual Investors: A Temporal Graph Network Approach with Mean-variance Efficient Sampling.
5th ACM Conference on AI in Finance (ICAIF 2024), New York, USA, November 2024, pp. 795-803. DOI: https://doi.org/10.1145/3677052.3698662

Abstract: Recommender systems can be helpful for individuals to make well-informed decisions in complex financial markets. While many studies have focused on predicting stock prices, even advanced models fall short of accurately forecasting them. Additionally, previous studies indicate that individual investors often disregard established investment theories, favoring their personal preferences instead. This presents a challenge for stock recommendation systems, which must not only provide strong investment performance but also respect these individual preferences. To create effective stock recommender systems, three critical elements must be incorporated: 1) individual preferences, 2) portfolio diversification, and 3) the temporal dynamics of the first two. In response, we propose a new model, Portfolio Temporal Graph Network Recommender (PfoTGNRec), which can handle time-varying collaborative signals and incorporates diversification-enhancing sampling. On real-world individual trading data, our approach demonstrates superior performance compared to state-of-the-art baselines, including cutting-edge dynamic embedding models and existing stock recommendation models. Indeed, we show that PfoTGNRec is an effective solution that can balance customer preferences with the need to suggest portfolios with high Return-on-Investment.
Keywords: recommender systems, financial recommendation, temporal graph network.
Javier Sanz-Cruzado, Edward Richards, Richard McCreadie. FAR-AI: A Modular Platform for Investment Recommendation in the Financial Domain.
46th European Conference on Information Retrieval (ECIR 2024), Demo track, Glasgow, Scotland, March 2024, pp. 267-271 (Volume 5). DOI: https://doi.org/10.1007/978-3-031-56069-9_30

Abstract: Financial asset recommendation (FAR) is an emerging sub-domain of the wider recommendation field that is concerned with recommending suitable financial assets to customers, with the expectation that those customers will invest capital into a subset of those assets. FAR is a particularly interesting sub-domain to explore, as unlike traditional movie or product recommendation, FAR solutions need to analyse and learn from a combination of time-series pricing data, company fundamentals, social signals and world events, relating the patterns observed to multi-faceted customer representations comprising profiling information, expectations and past investments. In this demo we will present a modular FAR platform; referred to as FAR-AI, with the goal of raising awareness and building a community around this emerging domain, as well as illustrate the challenges, design considerations and new research directions that FAR offers. The demo will comprise two components: 1) we will present the architecture of FAR-AI to attendees, to enable them to understand the how’s and the why’s of developing a FAR system; and 2) a live demonstration of FAR-AI as a customer-facing product, highlighting the differences in functionality between FAR solutions and traditional recommendation scenarios. The demo is supplemented by online-tutorial materials, to enable attendees new to this space to get practical experience with training FAR models.
Keywords: recommender systems, finance, machine learning.
Video: https://www.youtube.com/watch?v=omEyrkheaqg
Poster: https://javiersanzcruza.github.io/assets/posters/ecir2024-poster.pdf
Javier Sanz-Cruzado, Pablo Castells. RELISON: A Framework for Link Recommendation in Social Networks.
45th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2022). Resource track., Madrid, Spain, July 2022, pp. 2992–3002. DOI: https://doi.org/10.1145/3477495.3531730

Abstract: Link recommendation is an important and compelling problem at the intersection of recommender systems and online social networks. Given a user, link recommenders identify people in the platform the user might be interested in interacting with. We present RELISON, an extensible framework for running link recommendation experiments. The library provides a wide range of algorithms, along with tools for evaluating the produced recommendations. RELISON includes algorithms and metrics that consider the potential effect of recommendations on the properties of online social networks. For this reason, the library also implements network structure analysis metrics, community detection algorithms, and network diffusion simulation functionalities. The library code and documentation is available at https://github.com/ir-uam/RELISON.
Keywords: social networks, recommender systems, contact recommendation, link prediction.
Video: https://dl.acm.org/action/downloadSupplement?doi=10.1145%2F3477495.3531730&file=SIGIR22-rs1425.mp4
Poster: https://javiersanzcruza.github.io/assets/posters/sigir2022-poster.pdf
Code: https://github.com/ir-uam/RELISON
Javier Sanz-Cruzado, Craig Macdonald, Iadh Ounis, Pablo Castells. Axiomatic Analysis of Contact Recommendation Methods in Social Networks: An IR Perspective.
42nd European Conference on Information Retrieval (ECIR 2020), Lecture Notes in Computer Science, Vol. 12035. Online, April 2020, pp. 157-190. DOI: https://doi.org/10.1007/978-3-030-45439-5_12

Abstract: Contact recommendation is an important functionality in many social network scenarios including Twitter and Facebook, since they can help grow the social networks of users by suggesting, to a given user, people they might wish to follow. Recently, it has been shown that classical information retrieval (IR) weighting models – such as BM25 – can be adapted to effectively recommend new social contacts to a given user. However, the exact properties that make such adapted contact recommendation models effective at the task are as yet unknown. In this paper, inspired by new advances in the axiomatic theory of IR, we study the existing IR axioms for the contact recommendation task. Our theoretical analysis and empirical findings show that while the classical axioms related to term frequencies and term discrimination seem to havea positive impact on the recommendation effectiveness, those related to length normalization tend to be not desirable for the task.
Keywords: social networks, recommender systems, contact recommendation, information retrieval models, axiomatic information retrieval.
Slides: https://www.slideshare.net/JavierSanzCruzadoPui/ecir-2020-axiomatic-analysis-of-contact-recommendation-methods-in-social-networks-an-ir-perspective
Video: https://www.youtube.com/watch?v=wX6OaNPtadY&t=17539s
Code: https://github.com/ir-uam/contact-rec-axioms
Javier Sanz-Cruzado, Pablo Castells. A Simple Multi-Armed Nearest-Neighbor Bandit for Interactive Recommendation.
13th ACM Conference on Recommender Systems (RecSys 2019), Copenhagen, Denmark, September 2019, pp. 358-362. DOI: https://doi.org/10.1145/3298689.3347040

Abstract: Contact recommendation has become a common functionality in online social platforms, and an established research topic in the social networks and recommender systems fields. Predicting and recommending links has been mainly addressed to date as an accuracy-targeting problem. In this paper we put forward a different perspective, considering that correctly predicted links may not be all equally valuable. Contact recommendation brings an opportunity to drive the structural evolution of a social network towards desirable properties of the network as a whole, beyond the sum of the isolated gains for the individual users to whom recommendations are delivered - global properties that we may want to assess and promote as explicit recommendation targets. In this perspective, we research the definition of relevant diversity metrics drawing from social network analysis concepts, and linking to prior diversity notions in recommender systems. In particular, we elaborate on the notion of weak tie recommendation as a means to enhance the structural diversity of networks. In order to show the signification of the proposed metrics, we report experiments with Twitter data illustrating how state of the art contact recommendation methods compare in terms of our metrics; we examine the tradeoff with accuracy, and we show that diverse link recommendations result in a corresponding diversity enhancement in the flow of information through the network, with potential implications in mitigating filter bubbles.
Keywords: recommender systems, multi-armed bandits, Thompson sampling, interactive recommendation, k nearest neighbors.
Poster: https://javiersanzcruza.github.io/assets/posters/recsys2019-poster.pdf
Code: https://github.com/ir-uam/kNNBandit
Javier Sanz-Cruzado, Pablo Castells. Information Retrieval Models for Contact Recommendation in Social Networks.
41st European Conference on Information Retrieval (ECIR 2019), Lecture Notes in Computer Science, Vol. 11437. Cologne, Germany, April 2019, pp. 148-163. DOI: https://doi.org/10.1007/978-3-030-15712-8_10

Abstract: The fast growth and development of online social networks has posed new challenges for information retrieval and, as a particular case, recommender systems. A particularly compelling problem in this context is recommending network edges, that is, automatically predicting people that a given user may wish or benefit from connecting to in the network. This task has interesting particularities compared to more traditional recommendation domains, a salient one being that recommended items belong to the same space as the users they are recommended to. In this paper, we investigate the connection between the contact recommendation and the text retrieval tasks. Specifically, we research the adaptation of IR models for recommending contacts in social networks. We report experiments over data downloaded from Twitter where we observe that IR models, particularly BM25, are competitive compared to state-of-the art contact recommendation methods. We further find that IR models have additional advantages in computational efficiency, and allow for fast incremental updates of recommendations as the network grows.
Keywords: social networks, recommender systems, contact recommendation, information retrieval models, collaborative filtering.
Slides: https://www.slideshare.net/JavierSanzCruzadoPui/ecir-2019-information-retrieval-models-for-contact-recommendation-in-social-networks
Javier Sanz-Cruzado, Pablo Castells. Enhancing Structural Diversity in Social Networks by Recommending Weak Ties.
12th ACM Conference on Recommender Systems (RecSys 2018), Vancouver, Canada, October 2018, pp. 233-241. DOI: https://doi.org/10.1145/3240323.3240371

Abstract: Contact recommendation has become a common functionality in online social platforms, and an established research topic in the social networks and recommender systems fields. Predicting and recommending links has been mainly addressed to date as an accuracy-targeting problem. In this paper we put forward a different perspective, considering that correctly predicted links may not be all equally valuable. Contact recommendation brings an opportunity to drive the structural evolution of a social network towards desirable properties of the network as a whole, beyond the sum of the isolated gains for the individual users to whom recommendations are delivered - global properties that we may want to assess and promote as explicit recommendation targets. In this perspective, we research the definition of relevant diversity metrics drawing from social network analysis concepts, and linking to prior diversity notions in recommender systems. In particular, we elaborate on the notion of weak tie recommendation as a means to enhance the structural diversity of networks. In order to show the signification of the proposed metrics, we report experiments with Twitter data illustrating how state of the art contact recommendation methods compare in terms of our metrics; we examine the tradeoff with accuracy, and we show that diverse link recommendations result in a corresponding diversity enhancement in the flow of information through the network, with potential implications in mitigating filter bubbles.
Keywords: social networks, recommender systems, contact recommendation, diversity, novelty, redundancy, social network analysis, weak tie, evaluation, metric.
Slides: http://es.slideshare.net/JavierSanzCruzadoPui/recsys-2018-enhancing-structural-diversity-in-social-networks-by-recommending-weak-ties
Video: https://www.youtube.com/watch?v=jKMjVibQA2o&list=PLaZufLfJumb97F5iAcZ6nx6AWg6sy1cJ5&index=43
Poster: https://javiersanzcruza.github.io/assets/posters/recsys2018-poster.pdf
Javier Sanz-Cruzado, Sofía Marina Pepa, Pablo Castells. Recommending Contacts in Social Networks Using Information Retrieval Models.
5th Spanish Conference on Information Retrieval (CERI 2018), Zaragoza, Spain, June 2018, 19 pp. 1-8. DOI: https://doi.org/10.1145/3230599.3230619

Abstract: The fast growth and development of online social networks has posed new challenges for information retrieval and, as a particular case, recommender systems. A particularly compelling problem in this context is recommending network edges, that is, automatically predicting people that a given user may wish or benefit from connecting to in the network. This task has interesting particularities compared to more traditional recommendation domains, a salient one being that recommended items belong to the same space as the users they are recommended to. In this paper, we investigate the connection between the contact recommendation and the text retrieval tasks. Specifically, we research the adaptation of IR models for recommending contacts in social networks. We report experiments over data downloaded from Twitter where we observe that IR models, particularly BM25, are competitive compared to state-of-the art contact recommendation methods. We further find that IR models have additional advantages in computational efficiency, and allow for fast incremental updates of recommendations as the network grows.
Keywords: social networks, recommender systems, contact recommendation, information retrieval models, collaborative filtering.

Book chapters

Javier Sanz-Cruzado, Pablo Castells. Contact Recommendations in Social Networks.
Shlomo Berkovsky, Iván Cantador, Domonkos Tikk (Eds): Collaborative Recommendations: Algorithms, Practical Challenges and Applications, World Scientific Publishing, November 2018, pp. 519-570. DOI: https://doi.org/10.1142/9789813275355_0016

Abstract: The increasingly fast development and expansion of recommender systems technology over the last two and a half decades, along with the exponential growth of online social networks in the last few years, has given place to the concurrence of the two areas in several directions. The present chapter focuses on a specific area within this confluence: the recommendation of people to connect with in social networks. We analyze the specifics of contact suggestion as a very particular recommendation task, where both the target users and the target items are people. We give an overview of the most relevant state of the art algorithms in this area, including methods that were originally developed with slightly different problems in mind. We present a global empirical comparison of the reviewed algorithms in order to get a perspective of their comparative performance. We conclude discussing future possible directions for research and development in this area.
Keywords: recommender systems, social networks, contact recommendation, evaluation, novelty, diversity.

Workshop publications

Lubingzhi Guo, Javier Sanz-Cruzado, Richard McCreadie. University of Glasgow at the FinLLM Challenge Task: Adapting Llama for Financial News Abstractive Summarization.
Joint Workshop of the 8th Financial Technology and Natural Language Processing (FinNLP) and the 1st Agent AI for Scenario Planning (AgentScen) co-located with 33rd International Joint Conference on Artificial Intelligence (IJCAI 2024). Jeju, Republic of Korea, August 2024, pp. 127-132.

Abstract: In this paper, we explore different approaches for aligning Large Language Models (LLMs) with the objectives of the financial abstractive summarization shared task. This shared task focuses on using LLM to abstract news into concise summaries. We investigate three common strategies: few-shot learning, fine-tuning, and reinforcement learning, to adapt LLMs for this purpose, with the fine-tuned model ranked first on the leaderboard.
Keywords: finance, summarization, large language models.
Slides: https://javiersanzcruza.github.io/assets/slides/finnlp-2024.pdf
Youngbin Lee, Yeyin Kim, Javier Sanz-Cruzado, Richard McCreadie, Yongjae Lee. Stock Recommendations for Individual Investors: A Temporal Graph Network Approach with Mean-variance Efficient Learning.
IJCAI 2024 Workshop on Recommender Systems in Finance (Fin-RecSys 2024) co-located with 33rd International Joint Conference on Artificial Intelligence (IJCAI 2024). Jeju, Republic of Korea, August 2024, .

Abstract: In complex financial markets, recommender systems can play a crucial role in empowering individuals to make informed decisions. Existing studies predominantly focus on price prediction, but even the most sophisticated models cannot accurately predict stock prices. Also, many studies show that most individual investors do not follow established investment theories because they have their own preferences. Hence, the tricky point in stock recommendation is that recommendations should give good investment performance but also should not ignore individual preferences. To develop effective stock recommender systems, it is essential to consider three key aspects: 1) individual preferences, 2) portfolio diversification, and 3) temporal aspect of both stock features and individual preferences. In response, we develop the temporal graph network approach with mean-variance efficient learning PfoTGNRec, which can handle time-varying collaborative signals and incorporates diversification-enhancing learning. As a result, our model demonstrated superior performance compared to various baselines, including cutting-edge dynamic embedding models and existing stock recommendation models, in a sense that our model exhibited good investment performance while maintaining competitive in capturing individual preferences.
Keywords: recommender systems, financial recommendation.
Slides: https://javiersanzcruza.github.io/assets/slides/finrecsys-2024-stocks.pdf
Lubingzhi Guo, Javier Sanz-Cruzado, Richard McCreadie. Comparing the Impact of Financial Knowledge Graphs from Financial Reports and Wikidata in Asset Recommendation.
IJCAI 2024 Workshop on Recommender Systems in Finance (Fin-RecSys 2024) co-located with 33rd International Joint Conference on Artificial Intelligence (IJCAI 2024). Jeju, Republic of Korea, August 2024, .

Abstract: Financial asset recommender (FAR) systems suggest investment assets to customers based on past market information. Many of these models choose those securities which they estimate to be more profitable for customers. Financial knowledge graphs (KGs)-- data structures containing information about assets and their relations to other involved entities (companies, people) -- have been one of the data sources exploited to drive asset selection. Although the construction of knowledge graphs from different sources (news, reports) has previously been investigated, there has been limited analysis of the effect these construction strategies have for FAR. In this work, we compare two different knowledge graphs representing U.S. stocks under a unified FAR framework: a knowledge graph crawled from a general knowledge base, Wikidata, and a knowledge graph built by extracting entities and relations from 10K financial reports using the GoLLIE open information extraction model. We show that integrating these KGs in FAR can lead up to 10.7% improvements in monthly ROI. However, the nature of these graphs makes algorithms prone to bias the recommendations towards different asset types.
Keywords: knowledge graphs, profitability prediction, financial recommendation, profitability.
Slides: https://javiersanzcruza.github.io/assets/slides/finrecsys-2024-comparing.pdf
Javier Sanz-Cruzado, Nikolaos Droukas, Richard McCreadie. FAR-Trans: An Investment Dataset for Financial Asset Recommendation.
IJCAI 2024 Workshop on Recommender Systems in Finance (Fin-RecSys 2024) co-located with 33rd International Joint Conference on Artificial Intelligence (IJCAI 2024). Jeju, Republic of Korea, August 2024, . DOI: https://doi.org/10.48550/arXiv.2407.08692

Abstract: Financial asset recommendation (FAR) is a sub-domain of recommender systems which identifies useful financial securities for investors, with the expectation that they will invest capital on the recommended assets. FAR solutions analyse and learn from multiple data sources, including time series pricing data, customer profile information and expectations, as well as past investments. However, most models have been developed over proprietary datasets, making a comparison over a common benchmark impossible. In this paper, we aim to solve this problem by introducing FAR-Trans, the first public dataset for FAR, containing pricing information and retail investor transactions acquired from a large European financial institution. We also provide a bench-marking comparison between eleven FAR algorithms over the data for use as future baselines
Keywords: recommender systems, financial recommendation, profitability, interactions, time series, dataset.
Slides: https://javiersanzcruza.github.io/assets/slides/finrecsys-2024-fartrans.pdf
Javier Sanz-Cruzado, Richard McCreadie, Nikolaos Droukas, Craig Macdonald, Iadh Ounis. On Transaction-Based Metrics as a Proxy for Profitability of Financial Asset Recommendations.
3rd Workshop on Personalization & Recommender Systems in Financial Services (FinRec 2022) co-located with 16th ACM Conference on Recommender Systems (RecSys 2022). Seattle, USA, September 2022, .

Abstract: The use of recommender systems to assist in the provision of financial asset and portfolio recommendations to investors is increasing, spanning a wide range of algorithms and techniques. Several strategies have been devised for the evaluation of financial asset recommendations, with the two most prominent strategies measuring (a) the money customers could obtain if they followed the recommendations (profitability-based evaluation) and (b) the ability of models to predict future customer investments (transaction-based evaluation). If customers are effective investors, we would expect these two perspectives to be positively correlated. In this paper, we perform experiments over a new large-scale financial recommendation dataset with real customer investment transactions to validate this assumption. Surprisingly, we find that transaction and profitability-based metrics are in fact negatively correlated and moreover, algorithms that actively try to learn from past customer transactions lose money over the mid-term. A thorough analysis of model performance and customer transaction patterns over time illustrates that this is due to a set of confounding factors, namely: customers failing to beat the market with their investments; a tendency for the customers to favour different investment lengths; and the impact of global events such as the Covid-19 pandemic.
Keywords: recommender systems, financial recommendation, evaluation, metric, profitability, interactions.
Slides: https://javiersanzcruza.github.io/assets/slides/finrec2022.pdf
Video: https://www.youtube.com/watch?v=kIdgo6H5ems
Javier Sanz-Cruzado, Pablo Castells. Beyond Accuracy in Link Prediction.
3rd Workshop on Social Media for Personalization and Search (SoMePeAS 2019) co-located with 41st European Conference on Information Retrieval (ECIR 2019). Communications in Computer and Information Science , Vol. 1245. Cologne, Germany, April 2019, pp. 79-94. DOI: https://doi.org/10.1007/978-3-030-52485-2_9

Abstract: Link prediction has mainly been addressed as an accuracy-targeting problem in social network analysis. We discuss different perspectives on the problem considering other dimensions and effects that the link prediction methods may have on the network where they are applied. Specifically, we consider the structural effects the methods can have if the predicted links are added to the network. We consider further utility dimensions beyond prediction accuracy, namely novelty and diversity. We adapt specific metrics from social network analysis, recommender systems and information retrieval, and we empirically observe the effect of a set of link prediction algorithms over Twitter data.
Keywords: social networks, recommender systems, contact recommendation, diversity, novelty, redundancy, social network analysis, weak tie, evaluation, metric.
Javier Sanz-Cruzado, Sofía Marina Pepa, Pablo Castells. Structural Novelty and Diversity in Link Prediction.
9th International Workshop on Modeling Social Media (MSM 2018) co-located with The Web Conference 2018 (WWW 2018). Companion of The Web Conference 2018 . Lyon, France, April 2018, pp. 1347-1351. DOI: https://doi.org/10.1145/3184558.3191576

Abstract: Link prediction has mainly been addressed as an accuracy-targeting problem in the social networks field. We discuss different perspectives on the problem considering other dimensions and effects that the link prediction methods may have on the social network where they are applied. Specifically, we consider the structural effects the prediction can have if the predicted links are added to the network. We consider further utility dimensions beyond prediction accuracy, namely novelty and diversity. We discuss the adaptation, for this purpose, of specific network, novelty and diversity metrics from social network analysis, recommender systems, and information retrieval.
Keywords: social networks, recommender systems, contact recommendation, diversity, novelty, redundancy, social network analysis, weak tie, evaluation, metric.