Tomek Strzalkowski

Verified

Tomek verified their affiliation via an institutional email.

Learn more

Verified

Tomek verified their affiliation via an institutional email.

Learn more

• PhD
• Professor (Full) at Rensselaer Polytechnic Institute

Large language models (LLMs) have demonstrated impressive performance in mathematical and commonsense reasoning tasks using chain-of-thought (CoT) prompting techniques. But can they perform emotional reasoning by concatenating `Let's think step-by-step' to the input prompt? In this paper we investigate this question along with introducing a novel a...

The identification of Figurative Language (FL) features in text is crucial for various Natural Language Processing (NLP) tasks, where understanding of the author's intended meaning and its nuances is key for successful communication. At the same time, the use of a specific blend of various FL forms most accurately reflects a writer's style, rather...

Imageability is a psycholinguistic property of words that indicates how quickly and easily a word evokes a mental image or other sensory experience. Highly imageable words are easier to read and comprehend, and, as a result, their use in communications, such as social media, makes messages more memorable, and, potentially, more impactful and influe...

Efficient evaluation of dialogue agents is a major problem in conversational AI, with current research still relying largely on human studies for method validation. Recently, there has been a trend toward the use of automatic self-play and bot-bot evaluation as an approximation for human ratings of conversational systems. Such methods promise to al...

Intelligence can be understood as the timely delivery of actionable information. Our Cognitive Immersive Room for Intelligence Analysis Scenarios (CIRAS) supports foraging and processing information during time-critical scenarios. Intelligence has an ambiguous meaning and could either refer to the ability to learn and reason well using a logical ap...

Recent advances in large-scale language modeling and generation have enabled the creation of dialogue agents that exhibit human-like responses in a wide range of conversational scenarios spanning a diverse set of tasks, from general chit-chat to focused goal-oriented discourse. While these agents excel at generating high-quality responses that are...

No PDF available ABSTRACT Dialogue systems have become a popular research medium as recent advances in task-oriented and open-domain systems combined with deep learning technologies have increased the potential for practical applications across many disciplines. One such vein of applications involves multi-modal dialogue systems deployed in interac...

Understanding how humans respond to an ongoing pandemic and interventions is crucial to monitoring and forecasting the dynamics of viral transmission. Heterogeneous response over time and geographical regions may depend on the individual beliefs and information consumption patterns of populations. To address the need for more precise and accurate e...

In this article, we describe our method of modeling sociolinguistic behaviors of players in massively multi-player online games. The focus of this paper is leadership, as it is manifested by the participants engaged in discussion, and the automated modeling of this complex behavior in virtual worlds. We first approach the research question of model...

We present a generalized framework for domain-specialized stance detection, focusing on Covid-19 as a use case. We define a stance as a predicate-argument structure (combination of an action and its participants) in a simplified one-argument format, e.g., wear(a mask), coupled with a task-specific belief category representing the purpose (e.g., pro...

Achieving true human-like ability to conduct a conversation remains an elusive goal for open-ended dialogue systems. We posit this is because extant approaches towards natural language generation (NLG) are typically construed as end-to-end architectures that do not adequately model human generation processes. To investigate, we decouple generation...

We describe Panacea, a system that supports natural language processing (NLP) components for active defenses against social engineering attacks. We deploy a pipeline of human language technology, including Ask and Framing Detection, Named Entity Recognition, Dialogue Engineering, and Sty-lometry. Panacea processes modern message formats through a p...

We present a paradigm for extensible lexicon development based on Lexical Conceptual Structure to support social engineering detection and response generation. We leverage the central notions of ask (elicitation of behaviors such as providing access to money) and framing (risk/reward implied by the ask). We demonstrate improvements in ask/framing d...

We describe Panacea, a system that supports natural language processing (NLP) components for active defenses against social engineering attacks. We deploy a pipeline of human language technology, including Ask and Framing Detection, Named Entity Recognition, Dialogue Engineering, and Stylometry. Panacea processes modern message formats through a pl...

Social engineers attempt to manipulate users into undertaking actions such as downloading malware by clicking links or providing access to money or sensitive information. Natural language processing, computational sociolinguistics, and media-specific structural clues provide a means for detecting both the ask (e.g., buy gift card) and the risk/rewa...

Social engineers attempt to manipulate users into undertaking actions such as downloading malware by clicking links or providing access to money or sensitive information. Natural language processing, computational sociolinguistics, and media-specific structural clues provide a means for detecting both the ask (e.g., buy gift card) and the risk/rewa...

A common approach, adopted by most current research, represents users of a social media platform as nodes in a network, connected by various types of links indicating the different kinds of inter-user relationships and interactions. However, social media dynamics and the observed behavioral phenomena do not conform to this user-node-centric view, p...

Social engineering attacks are a significant cybersecurity threat putting individuals and organizations at risk. Detection techniques based on metadata have been used to block such attacks, but early detection success is minimal. Natural language processing and computational sociolinguistics techniques can provide a means for detecting and counteri...

GitHub is a popular source code hosting and development service that supports distributed teams working on large and small software projects, particularly open-source projects. According to Wikipedia, as of April 2017 GitHub supports more than 20 million users and more than 57 million repositories. In addition to version control and code updates fu...

That games can be used to teach specific content has been demonstrated numerous times. However, although specific game features have been conjectured to have an impact on learning outcomes, little empirical research exists on the impact of iterative design on learning outcomes. This article analyzes two games that have been developed to train an ad...

The engaging nature of video games has intrigued learning professionals attempting to capture and retain learners’ attention. Designing learning interventions that not only capture the learner’s attention, but also are designed around the natural cycle of attention will be vital for learning. This paper introduces the temporal attentive observation...

Educational games have generated attention for their potential to teach more successfully and with longer-lasting outcomes than those generated by traditional teaching methods. Questions remain, however, about what features of games enhance learning. This study investigates the effects of art style and narrative complexity on training outcomes of a...

This article discusses the design and development of two serious games intended to train people to reduce their reliance on cognitive biases in their decision-making in less than an hour each. In our development process, we found a tension between rich and flexible experimentation and exploration experiences and robust learning experiences that ens...

The development of a serious game combines the skills of numerous disciplines, from subject matter experts on the topic being taught; to story developers, game designers, and software developers; to instructional designers, educational assessment scientists, and others. This section provides commentary on the Intelligence Advanced Research Projects...

As research on serious games continues to grow, we investigate the efficacy of digital games to train enhanced decision making through understanding cognitive biases. This study investigates the ability of a 30-minute digital game as compared with a 30-minute video to teach people how to recognize and mitigate three cognitive biases: fundamental at...

In this article, we present a method to validate a multilingual (English, Spanish, Russian, and Farsi) corpus on imageability ratings automatically expanded from MRCPD (Liu et al., 2014). We employed the corpus (Brysbaert et al., 2014) on concreteness ratings for our English MRCPD+ validation because of lacking human assessed imageability ratings a...

In this article, we describe our method of modeling socio-linguistic behaviors of players in massively multi-player online games. The focus of this paper is leadership, as it is manifested by the participants engaged in discussion, and the automated modeling of this complex behavior in virtual worlds. We first approach the research question of mode...

Although human use of heuristics can result in 'fast and frugal' decision-making, those prepotent tendencies can also impair our ability to make optimal choices. Previous work had suggested such cognitive biases are resistant to mitigation training. Serious games offer a method to incorporate desirable elements into a training experience, and allow...

In this article, we outline a novel approach to the automated analysis of cross-cultural conflicts through the discovery and classification of the metaphors used by the protagonist parties involved in the conflict. We demonstrate the feasibility of this approach on a prototypi-cal conflict surrounding the appropriate management and oversight of gun...

The number of educational or serious games (SGs) available to educators has increased in recent years as the cost of game development has been reduced. A benefit of SGs is that they employ not only lesson content but also knowledge contexts where learners can connect information to its context of use with active participation and engagement. This,...

Drawing from recent research on the ability of video games to satisfy psychological needs, this paper identifies how the presence of rewards influences learning complex concepts and tasks using an educational video game. We designed and developed two 60-minute educational games with and without a range of reward features and examined learning outco...

Although considerable research has identified patterns in online communication and interaction related to a range of individual characteristics, analyses of age have been limited, especially those that compare age groups. Research that does examine online communication by age largely focuses on linguistic elements. However, social identity approach...

Cognitive biases are systematic errors that result from reliance on heuristics in decision-making. Such biases are typically automatic and unconscious influences on behavior, and can occur in a wide range of situations and contexts. Cognitive biases are generally resistant to mitigation training. This project adopted a novel approach to develop com...

Background Engagement has been identified as a crucial component of learning in games research. However, the conceptualization and operationalization of engagement vary widely in the literature. Many valuable approaches illuminate ways in which presence, flow, arousal, participation, and other concepts constitute or contribute to engagement. Howeve...

This article makes two contributions towards the use of lexical resources and corpora; specifically making use of them for gaining access to and using word associations. The direct application of our approach is for detecting linguistic and conceptual metaphors automatically in text. We describe our method of building conceptual spaces, that is, de...

This article describes a novel approach to automated determination of affect associ-ated with metaphorical language. Affect in language is understood to mean the at-titude toward a topic that a writer at-tempts to convey to the reader by using a particular metaphor. This affect, which we will classify as positive, negative or neutral with various d...

In this article, we present details about our ongoing work towards building a repository of Linguistic and Conceptual Metaphors. This resource is being developed as part of our research effort into the large-scale detection of metaphors from unrestricted text. We have stored a large amount of automatically extracted metaphors in American English, M...

Recent studies in metaphor extraction across several languages (Broadwell et al., 2013; Strzalkowski et al., 2013) have shown that word imageability ratings are highly correlated with the presence of metaphors in text. Information about imageability of words can be obtained from the MRC Psycholinguistic Database (MRCPD), which is a collection of hu...

Background. Engagement has been identified as a crucial component of learning in games research. However, the conceptualization and operationalization of engagement vary widely in the literature. Many valuable approaches illuminate ways in which presence, flow, arousal, participation, and other concepts constitute or contribute to engagement. Howev...

Educational games have proliferated, but questions remain about the effectiveness at teaching both in the short- and long-term. Also unclear is whether particular game features have positive effects on learning. To examine these issues, this paper describes a controlled experiment using an educational game that was professionally developed to teach...

This article describes our novel approach to the automated detection and analysis of meta-phors in text. We employ robust, quantitative language processing to implement a system prototype combined with sound social science methods for validation. We show results in 4 different languages and discuss how our methods are a significant step forward fro...

In this paper, we describe a novel approach to automatically detecting and tracking discus-sion dynamics in Internet social media by fo-cusing on attitude modeling of topics. We characterize each participant's attitude to-wards topics as Topical Positioning, employ Topical Positioning Map to represent the posi-tions of participants with respect to...

The reliable automated identification of metaphors still remains a challenge in metaphor research due to ambiguity between semantic and contextual interpretation of individual lexical items. In this article, we describe a novel approach to metaphor identification which is based on three intersecting methods: imageability, topic chaining, and semant...

In this article, we present a novel approach towards the detection and modeling of complex social phenomena in multiparty interactions, including leadership, influence, pursuit of power and group cohesion. We have developed a two-tier approach that relies on observable and computable linguistic features of conversational text to make predictions ab...

Semantic and syntactic features found in text can be used in combination to statistically predict linguistic devices such as hedges in online chat. Some features are better indicators than others, and there are cases when multiple features need to be considered together to be useful. Once the features are identified, it becomes an optimization prob...

In this article, we present a novel approach towards the detection and modeling of complex social phenomena in multi-party discourse, including leadership, influence, pursuit of power and group cohesion. We have developed a two-tier approach that relies on observable and computable linguistic features of conversational text to make predictions abou...

In this article, we present our novel approach towards the detection and modeling of complex social phenomena in multi-party discourse, including leadership, influence, pursuit of power and group cohesion. We have developed a two-tier approach that relies on observable and computable linguistic features of conversational text to make predictions ab...

Recent advances in automated analysis of on-line chat data allow us to draw conclusions about social behavior, such as leadership, in small groups previously possible only through manual methods of observation and analysis. We have applied such methods to comparable English and Chinese language data, defined a new language use called Tension Focus,...

In this paper, we describe a new approach to semi-supervised adaptive learning of event extraction from text. Given a set of examples and an un-annotated text corpus, the BEAR system (Bootstrapping Events And Relations) will automatically learn how to recognize and understand descriptions of complex semantic relationships in text, such as events in...

Research on group cohesion—defined via a set of features such as group unity, group performance, collective efficacy, and group norms—has primarily focused onparticipant interviews and surveys where group members report their own assessments of the group (Casey-Campbell & Martens, 2009; Demock & Devine, 1997). Less research is available on sociolin...

In this paper, we describe a novel approach to computational modeling and understanding of social and cultural phenomena in multi-party dialogues. We developed a two-tier approach in which we first detect and classify certain socio-linguistic behaviors (SLB), including topic control, disagreement, and involvement, that serve as first order models f...

Information retrieval (IR) involves retrieving information from stored data, through user queries or pre-formulated user profiles. The information can be in any format. IR typically advances over four broad stages viz., identification of text types, document preprocessing, document indexing, and query processing and matching the same to documents....

Classification of dialogue acts constitutes an integral part of various natural language processing applications. In this paper, we present an application of this task to Urdu language online multi-party discourse. With language specific modifications to established techniques such as permutation of word order in detected n-grams and variation of n...

This research is part of a larger project that involves developing computational tools to model and recognize communicative behavior in online environments. Specifically, the paper reports on a series of metrics which have been designed to reveal varying degrees of influence and involvement in online interactions.

A social robot is a robotic platform that supports natural in-teraction with people in a human-scale environment. Such a platform allows interesting opportunities for both tradi-tional Computer Science students and students from other disciplines, such as psychology, philosophy, design and com-munications. In this paper, we describe a new social ro...

We describe a novel approach to computational modeling and understanding of social and cultural phenomena in multi-party online dialogues. We developed a two-tier approach in which we first detect and classify social language uses (LU) in discourse, including topic control, task control, disagreement, and involvement. These languages uses are the s...

We describe an annotation tool developed to assist in the creation of multimodal action-communication corpora from on-line massively multi-player games, or MMGs. MMGs typically involve groups of players (5--30) who control their avatars, perform various activities (questing, competing, fighting, etc.) and communicate via chat or speech using assume...

We present in this paper, the application of a novel approach to computational modeling, understanding and detection of social phenomena in online multi-party discourse. A two-tiered approach was developed to detect a collection of social phenomena deployed by participants, such as topic control, task control, disagreement and involvement. We discu...

The purpose of this research was to advance the understanding of the behavior of small groups in online chat rooms. The research was conducted using Internet chat data collected through planned exercises with recruited participants. Analysis of the collected data led to construction of preliminary models of social behavior in online discourse. Some...

In this paper, we report our efforts in building a multi-lingual multi-party online chat corpus (MMPC) in order to develop a firm understanding in a set of social constructs such as agenda control, influence, and leadership as well as to computationally model such constructs in online interactions. These automated models will help capture the dialo...

In this paper, we describe a novel approach to computational modeling and understanding of social and cultural phenomena in multi-party dialogues. We developed a two-tier approach in which we first detect and classify certain sociolinguistic behaviors, including topic control, disagreement, and involvement, that serve as first-order models from whi...

We introduce COLLANE, an experimental collaborative analytic environment that allows a group of professional analysts to work together effectively on complex, multifaceted information problems. COLLANE has been developed to investigate innovative ways of harnessing the power of collaboration so that to maximize the quality of the analytical product...

We describe an interactive question answering system, HITIQA, which helps users find answers to complex analytical problems. Such problems often necessitate the user to submit not one but an entire series of questions, both simple and complex, and then to negotiate the final content and form of the answer. HITIQA advances research in human–computer...

We describe a procedure for quantitative evaluation of interactive question-answering systems and illustrate it with application to the High-Quality Interactive Question-Answering (HITIQA) system. Our objectives were (a) to design a method to realistically and reliably assess interactive question-answering systems by comparing the quality of report...

The purpose of this work is to identify potential evaluation criteria for interactive, analytical question-answering (QA) systems by analyzing evaluative comments made by users of such a system. Qualitative data collected from intelligence analysts during interviews and focus groups were analyzed to identify common themes related to performance, us...

In this paper, we discuss how to utilize the co-occurrence of answers in building an automatic question answering system that answers a series of questions on a specific topic in a batch mode. Experi- ments show that the answers to the many of the questions in the series usually have a high degree of co-occurrence in rele- vant document passages. T...

The authors report on a series of experiments to automate the assessment of document qualities such as depth and objectivity. The primary purpose is to develop a quality-sensitive functionality, orthogonal to relevance, to select documents for an interactive question-answering system. The study consisted of two stages. In the classifier constructio...

We present a natural-language customer service application for a telephone banking call center, developed as part of the Amitiés dialogue project (Automated Multilingual Interaction with Information and Services). Our dialogue system, based on empirical data gathered from real call-center conversations, features data-driven techniques that allow fo...

This year, we made changes to the passage/sentence retrieval component of ILQUA in handling factoid and list questions. All the other components remain same.

We describe an interactive approach to question answering where the user and the system first negotiate the scope and shape of information being sought and then cooperate in locating and assembling the answer. The system, which we call HITIQA11, has access to a large repository of unprocessed and unformatted data, and is additionally equipped with...

Automated question answering - the ability of a machine to answer questions, simple or complex, posed in ordinary human language - is one of today’s most exciting technological developments. It has all the markings of a disruptive technology, one that is poised to displace the existing search methods and establish new standards for user-centered ac...

We analyzed textual properties of documents to identify predictive variables for various document qualities by means of statistical and linguistic methods. We have created a collection of 1000 documents, each document has been judged in terms of nine document qualities (accuracy, reliability, objectivity, depth, author/producer credibility, readabi...

In this paper we report preliminary results of a study to develop, and subsequently to automate, new metrics for assessment of information quality in text documents, particularly in news. Through focus group studies, quality judgment experiments, and textual feature extraction and analysis, we were able to generate nine quality aspects and apply th...

The work reports some initial success in extending the Rutgers Paradigm of IR evaluation to the realm of concrete measurement, not in information retrieval per se, but in the arguably more complex domain of Question Answering. Crucial to the paradigm are two components: cross evaluation, and an analytical model that controls for the potential probl...

The goal of this research is to automatically predict human judgments of document qualities such as subjectivity, verbosity and depth. In this paper, we explore the behavior of adjectives as indicators of subjectivity in documents. Specifically, we test whether a subset of automatically derived subjective adjectives (Wiebe, 2000b), selected a prior...

In addition to relevance, there are other factors that contribute to the utility of a document. For examples, content properties like depth of analysis and multiplicity of viewpoints, and presentational properties like readability and verbosity, all will affect the usefulness of a document. These kinds of relevance-independent properties are diffic...