SCORE prototype accepted to AAAI 2022 Demonstrations Program
22 October 2021
"A Synthetic Prediction Market for Estimating Confidence in Published work" is accepted to the demonstrations program at AAAI 2022. The video submission showcases our prototype system, built for our SCORE program, which estimates confidence in published scholarly work in the social and behavioral sciences. This work lays the foundation for a research agenda that creatively uses AI for peer review.
Paper accepted to NeurIPS Workshop on Algorithmic Fairness through the Lens of Causality and Robustness
21 October 2021
"Structural Interventions on Automated Decision Making Systems" led by postdoc Efren Cruz Cortes with collaborator Debashis Ghosh, is accepted to NeurIPS AFCR 2021. The paper engages in a systemic analysis, aided by structural causal models, to compare social interventions to algorithmic interventions in automated decision system to identify bias outside the algorithmic stage and propose joint interventions on social dynamics and algorithm design.
Paper accepted to Network Neuroscience
16 September 2021
"Feeding the machine: challenges to reproducible predictive modeling in resting-state connectomics" led by graduate student Andrew Cwiek with collaborators Brad Wybe, Vasant Honavar and Frank Hillary, is accepted to Network Neuroscience. The paper is a review of the use of ML in the network neurosciences and provides recommendations to overcome common methodological pitfalls. A preprint is available here.
Paper accepted to Results in Control and Optimization
13 September 2021
"Design and Analysis of a Synthetic Prediction Market using Dynamic Convex Sets" led by graduate students Nishanth Nakshatri and Arjun Menon, with collaborators Lee Giles and Christopher Griffin, has been accepted to Results in Control and Optimization. The work outlines the theoretical underpinnings for the synthetic prediction market we are developing as part of our DARPA SCORE effort. The paper is available here.
Join us @Metascience 2021
9 September 2021
We'll be discussing our DARPA SCORE effort to develop AI for evaluating confidence in published research claims in the social and behavioral sciences. More here! Conference starts next Thursday -- exciting program all around!
Lead article in First Monday explores self-disclosure during COVID
1 July 2021
"A study of self-disclosure during the Coronavirus pandemic" is led by graduate student Taylor Blose and analyzes instances of self-disclosure over the last year highlighting a massive spike in personal information sharing at critical moments in the pandemic timeline. Find it here!
Paper accepted to ECML-PKDD 2021
29 June 2021
"Self-disclosure on Twitter during COVID-19 Pandemic: A Network Perspective" is led by graduate student Prasanna Umar and explores the phenomenon of self-privacy violations during the COVID pandemic from a network theoretic lens.
New award from AFOSR on cognitively-plausible artificial agents
23 June 2021
We've received funding with Dr. Alan Wagner for a new 3-year project entitled, "Toward Cognitive Realism in Game Theoretic Models of Social Behavior." The project will computationally represent Construal Level Theory to develop agents that reason over representations of people, places and things with varying levels of abstraction. The work will offer new perspective on experience-based reasoning and complement current efforts in reinforcement learning. *Actively seeking a graduate student for this new work.
Paper introducing our DARPA SCORE collaboration up on SocArXiv
04 May 2021
The paper introduces the massive-scale effort to create and validate algorithms to provide confidence scores for research claims. See it here: https://osf.io/preprints/socarxiv/46mnb
Paper accepted to Sci-K @ The Web Conference 2021
17 February 2021
"Extraction and evaluation of statistical information from social and behavioral science papers" has been accepted to the Workshop on Scientific Knowledge at The Web Conference. The work presents a pipeline for extraction of statistical information (p-vals, sample size, number of hypotheses tested) from full-text scientific documents.
Paper accepted to PoPETS 2021
01 February 2021
Work with Dr. Shomir Wilson and graduate student Jooyoung Lee has been accepted to the Privacy Enhancing Technologies Symposium (PoPETS) 2021. "Digital inequality through the lens of self-disclosure" examines relationships between U.S.-based Twitter users' socio-demographic characteristics and their privacy behaviors.
Paper accepted to SDU @ AAAI-21
08 December 2020
"Understanding and predicting retractions of published work" has been accepted to the AAAI-21 Workshop on Scientific Document Understanding. The work develops a classifier to separate a set of retracted papers in the social and behavioral sciences with a comparable non-retracted set. The work is part of our DARPA SCORE effort and was led by graduate student Ajay Modukuri.
Two papers accepted to IEEE ISTAS
10 November 2020
Work with Dr. Caitlin Grady and graduate student Lauren Dennis has been accepted to IEEE's Symposium on Technology and Society (ISTAS). "When smart systems fail: the ethics of cyber-physical critical infrastructure risk" and "Analyzing cyber-physical threats to Pennsylvania dams through a lens of vulnerability" will be presented at the conference later this week. Both papers explore ethical implications of increasingly interdependent cyber-physical critical infrastructure.
Dr. Rajtmajer and colleagues receive award from the NSF's National Center for Science and Engineering Statistics (NCSES)
03 November 2020
Dr. Rajtmajer and colleagues at Quantitative Scientific Solutions, LLC have received funding from the National Science Foundation's NCSES to study trends in research and publishing, integrating data from the NSF's Survey of Doctorate Recipients. The broader aim is of the project is data-driven policies on scientific processes and funding, with specific focus on underrepresented PhD recipients.
Paper accepted to IEEE Big Data
21 October 2020
"A Study of Self-Privacy Violations in Online Public Discourse" will appear at IEEE Big Data 2020. The work develops a supervised approach to detect instances of self-disclosure in public discourse, and explores when and how these behaviors conform to group norms. The work is led by graduate student Prasanna Umar and Anna Squicciarini.
Paper accepted to Findings of EMNLP
22 September 2020
Our paper, "A Semantics-based Approach to Disclosure Classification in User-Generated Online Content", will appear in Findings of EMNLP. The work presents an approach to detect emotional and information self-disclosure in natural language through the use of Semantic Role Labeling. The work is led by graduate student Chandan Akiti.
Paper accepted to Applied Network Science
21 August 2020
"Fragility of a multilayer network of international supply chains" has been accepted for publication in Applied Network Science. The paper explores propagation of economic shocks along intranational supply chains through network cascade modeling over multiregional input-output linkages. A measure of fragility is proposed, and economic sectors disrupted by COVID-19 are highlighted.
Dr. Rajtmajer and colleagues receive award from the NSF's RAPID program
27 May 2020
Dr. Rajtmajer and Dr. Anna Squicciarini have received funding through the National Science Foundation's RAPID program to study privacy risks emergent with online behavior during the Coronavirus pandemic. The work aims to better understand how trust is established in online settings and whether that process is expedited during crisis.
Dr. Rajtmajer and colleagues receive award from the Office of Naval Research
15 May 2020
Dr. Rajtmajer, Dr. Aiping Xiong, Dr. Christopher Griffin, and our collaborators at Quantitative Scientific Solutions (QS-2) have received funding through the Navy's STTR program to study and mitigate botnets on Twitter. The project, Sociolinguistic Information Filtering Tool (SIFT), leverages novel natural language processing, network science and game theoretic approaches to characterize classes of malicious accounts and provide users with more sophisticated options to filter this content.
Dr. Rajtmajer, Dr. Vasant Honavar, Dr. Daniel Susser, and Dr. Jose Soto receive award from the Institute for Computational and Data Sciences
07 May 2020
Dr. Rajtmajer and colleagues have been awarded seed funds from Penn State's Institute for Computational and Data Sciences for work to empirically evaluate causal inference-based tests of algorithmic fairness. These studies will inform new work on fairness and discrimination in AI.
New work on privacy during Coronavirus released on arXiv
21 April 2020
Dr. Rajtmajer and colleagues have shared a new study of self-disclosure on Twitter during the Coronavirus pandemic. We suggest that increased disclosure may serve support-seeking during crisis. A preprint is available here.
Paper on Cognitive Security accepted to the ACM Symposium on the Science of Security (HoTSoS)
22 February 2020
"Automated Influence and the Challenge of Cognitive Security" has been accepted to ACM's Symposium on the Science of Security (HoTSoS 2020). In this work, Dr. Rajtmajer and Dr. Daniel Susser consider the threat of computational propaganda to national security and the ethical questions it raises.
New work on Russian Twitter Operations released on arXiv
27 January 2020
Dr. Rajtmajer and colleagues have shared new work revealing coordination in Russian influence operations on Twitter. The work uses a dynamical systems model to construct families of users with common harmonics and develops a taxonomy of strategic behavior. A preprint is available here.
Team to compete in AAAI shared task on affect understanding (AffCon 2020)
16 January 2020
Work led by Masters student Chandan Akiti will be presented at AffCon 2020 in February 2020. The team will compete in the Computational Linguistics Affect Understanding (Cl-Aff) Shared Task on modeling interactive affective responses. The team has developed a novel approach for the joint classification of self-disclosure and supportiveness in short text, leveraging BERT, LSTM and CNN neural networks.
Dr. Rajtmajer and Dr. Shomir Wilson receive accelerator award for work on privacy and socioeconomic factors
03 January 2020
Dr. Rajtmajer and Dr. Shomir Wilson have been awarded seed funds from Penn State's Center for Social Data Analytics (C-SoDA) for joint work exploring the effects of socioeconomic factors on privacy behaviors online. Proposed work is motivated by emerging evidence that current practices for informing technology users about privacy disproportionately fail to help the very populations that are most vulnerable.
Team selected for DARPA's SCORE program
23 October 2019
Dr. Rajtmajer and colleagues (Penn State, Texas A&M, Old Dominion, Microsoft Research) have been selected by the Defense Advanced Research Projects Agency (DARPA) to develop AI to score the credibility of research claims in the social and behavioral sciences. The team will build synthetic prediction markets populated by artificial agents (bot traders) that reason on information extracted from published papers and metadata. See the article in the Penn State News.
Paper accepted to IEEE Big Data 2019
17 October 2019
"Toward Image Privacy Classification and Spatial Attribution of Private Content" has been accepted to IEEE Big Data 2019. The work extends the problem of determining a single privacy label for a given image to jointly inferring a privacy label and detecting the specific areas of sensitive content within a privately labeled image. The paper is joint work with Drs. Haoti Zhong, Anna Squicciarini and David Miller.
Paper accepted to Complex Networks 2019
7 October 2019
"Performance of a Multiplex Commodity Flow Network in the United States Under Disturbance" has been accepted to Complex Networks 2019. The work furthers understanding of supply chain and commodity trade networks through construction and analysis of the interstate input-out multiplex network of the U.S. commodity and service sectors. The paper is joint work with Dr. Caitlin Grady (Penn State) and Dr. Alfonso Mejia (Penn State).
Paper accepted to CIKM 2019
9 August 2019
"Rating Mechanisms for Sustainability of Crowdsourcing Platforms" has been accepted to CIKM 2019. The paper, in collaboration with Dr. Chexi Qiu (Rowan University) and Dr. Anna Squicciarini (Penn State) introduces rating mechanisms to evaluate the behavior of task requesters in crowdsourcing platforms. We take a game theoretic approach, validated on data from Amazon Mechanical Turk.
Paper accepted to GAMESEC 2019
28 July 2019
"Power Law Public Goods Game for Personal Information Sharing in News Comments" has been accepted to GAMESEC 2019. The work proposes a public goods game model of user sharing in an online commenting forum. In particular, we assume that users who share personal information incur an information cost but reap the benefits of a more extensive social interaction. Freeloaders benefit from the same social interaction but do not share personal information. A preprint of the paper is available on the arXiv: https://arxiv.org/abs/1906.01677.
Dr. Rajtmajer and Dr. Anna Squicciarini awarded College of IST seed grant
13 June 2019
Dr. Rajtmajer and Dr. Anna Squicciarini are amongst eight recipients of funding from the College of IST’s Seed Grant Program. The project, entitled "Game-theoretic Modeling of Individual Self-disclosure Online" aims to formalize models of information sharing and privacy behaviors in public online environments.
Dr. Rajtmajer and Dr. Ben Johnson receive award from Institute for Cyberscience
16 April 2019
Dr. Rajtmajer and Dr. Ben Johnson have received seed funding from Penn State's Institute for Cyberscience for proposed work, "Leveraging AI for Game-Theoretic Models of Judicial Decision Making". The project will explore rich historical data to better understand and model the mechanisms and impacts of judicial decision making.
Rock Ethics ETE Fellows to present at undergraduate research exhibition
15 April 2019
The Rock Ethics Institute's Engineering, Technology and Ethics fellows (Jack Iffert, Ryan Jacobs, Vidur Mishra, Christopher Thomas and Nicholas Wanner) will present their work, titled "Security Risks and Ethical Implications for the Automation of Commercial Port Facilities" at the 2019 Penn State Undergraduate Research exhibition. Their work is the culmination of a year-long project, co-advised by Dr. Rajtmajer, Dr. Caitlin Grady (Civil Engineering), Dr. Daniel Susser (IST) and Dr. Alan Wagner (Aerospace Engineering).
Dr. Rajtmajer to moderate panel on security and emerging technologies
27 March 2019
Dr. Rajtmajer will moderate a panel discussion on Autonomous Systems and Emerging Technologies at the Center for Security Research and Education's Security and the Autonomous Future symposium, April 4th and 5th. The event will bring together national experts to explore how autonomous systems will effect our personal, national and global security.
Paper accepted to SIAM Journal on Applied Dynamical Systems
14 February 2019
Work with Dr. Christopher Griffin, Dr. Anna Squicciarini and Dr. Andrew Belmonte has been accepted to SIAM Journal on Applied Dynamical Systems (SIADS). The work, entitled "Consensus and Information Cascades in Game-Theoretic Imitation Dynamics with Static and Dynamic Network Topologies" provides a game-theoretic framework for modeling information spread online. A preprint is available on the arXiv: https://arxiv.org/abs/1903.11429.
Paper accepted to WWW 2019
21 January 2019
Work with Prasanna Umar and Dr. Anna Squicciarini has been accepted as a short paper to The Web Conference (WWW) 2019. The work, entitled "Detection and Analysis of Self-Disclosure in Online News Commentaries" presents a novel approach to the detection of language indicative of various types of self-disclosure, leveraging both syntactic and semantic information present in text.
Dr. Rajtmajer and Dr. Daniel Susser receive award from Center for Security Research and Education
20 January 2019
Dr. Rajtmajer and Dr. Daniel Susser have been awarded seed funding from Penn State's Center for Security Research and Education. The project, "Understanding the Impact of Self-Disclosure on Cognitive Security", will explore the contexts which prime individuals to self-disclose online in order to better understand the potential weaponization of personal information. The 1-year effort will engage graduate and undergraduate students from the College of IST, the Department of Philosophy, and across campus.
Beyond the Crisis: Research Ethics in the Age of Open Science
26 September 2018
Dr. Rajtmajer will present "Beyond the Crisis: Research Ethics in the Age of Open Science" at the Rock Ethics Institute Colloquia Series on October 4, 2018. The talk with discuss recent progress toward an interpretable scholarly record through promotion of transparency and openness in reporting and publishing.