In cognitive systems and cognitive linguistics, primitive decomposition systems attempt to explain cognitive phenomena by breaking things down into conceptual building blocks and provide rich and flexible representations for systems. A prime example is the Schank-Minsky Conceptual Dependency Trans-frames system, which maintains a commitment to keeping the number of primitives small and allowing them to be combined in complex ways in representing meaning, knowledge, and dynamic episodic memory. Motivated by the desire to keep the set of primitives small, this paper describes an effort to eliminate the Conceptual Dependency INGEST primitive and reconstitute its uses through combinations of the CD PTRANS primitive and CD’s representations of containment. The implementation is performed in Babel, an automated paraphrase generation system which generates English realizations of CD structures and which has been used in multiple natural language understanding and story understanding systems. The implementation combines the discrimination nets used for selecting word senses for the INGEST primitive with those for the PTRANS primitive. Once the implementation was complete, we also ran Babel using the new structures to generate paraphrases of CD structures and to determine the degree of success in our primitive re-expression endeavor.

Narrative story generation has gained emerging interest in the field of large language models. The present paper aims to compare stories generated by an LLM only (non-interleaved) with those generated by interleaving human-generated and LLM-generated text (interleaved). The study’s hypothesis is that interleaved stories would perform better than non-interleaved stories. To verify this hypothesis, we conducted two tests with roughly 500 participants each. Participants were asked to rate stories of each type, including an overall score or preference and four facets—logical soundness, plausibility, understandability, and novelty. Our findings indicate that interleaved stories were in fact less preferred than non-interleaved stories. The result has implications for the design and implementation of our story generators. This study contributes new insights into the potential uses and restrictions of interleaved and non-interleaved systems regarding generating narrative stories, which may help to improve the performance of such story generators.

One of the hallmarks of the Schank-Minsky Conceptual Dependency Trans-Frames meaning representation system is that it attempts to express complex meanings by building large and complex conceptual structures using a relatively small number of primitives. Recently comparisons of image schemas with Conceptual Dependency primitives revealed ways of possibly reducing the number of primitives while maintaining the expressiveness of the set—an important research goal because it increases the flexibility and richness of the primitive-decomposed structures in a way that better approximates human cognition. Inspired by this prior work, we employ a paraphrase generation system to explore the replacement of the Conceptual Dependency INGEST primitive by PTRANS and CONTAIN , which are the analogs of Source_Path_Goal and Containment image schemas. The results of the study bring us a step closer to a possible unification of image schemas with Conceptual Dependency.

In this work, we are concerned with developing cognitive representations that may enhance the ability for self-supervised learning systems to learn language as part of their world explorations. We apply insights from in-depth language understanding systems to the problem, specifically representations which decompose language inputs into language-free structures that are complex combinations of primitives representing cognitive abstractions such as object permanence, movement, and spatial relationships. These decompositions, performed by a system traditionally called a conceptual analyzer, link words with complex non-linguistic structures that engender the rich relations between language expressions and world exploration that are a familiar aspect of intelligence. We focus on improving and extending both the Conceptual Dependency (CD) representation system, its primitive decompositions, and its conceptual analyzer, choosing as our corpus the ProPara ("Process Paragraphs") dataset, which consists of paragraphs describing biological, chemical, and physical processes of the kind that appear in gradeschool science textbooks (e.g., photosynthesis, erosion). In doing so, we avoid the significant challenges of decomposing concepts involving communication, thought, and complex social interactions. To meet the challenges of this dataset, we contribute a mental motion pictures representation system with important innovations, such as using image schemas in place of CD primitives and decoupling containment relationships into separate primitives.

Describing the content of a visual image is a fundamental ability of human vision and language systems. Over the past several years, researchers have published on major improvements on image captioning, largely due to the development of deep learning systems trained on large data sets of images and human-written captions. However, these systems have major limitations, and their development has been narrowly focused on improving scores on relatively simple “bag-of-words” metrics. Very little work has examined the overall complex patterns of the language produced by image-captioning systems and how it compares to captions written by humans. In this paper, we closely examine patterns in machine-generated captions and characterize how conventional metrics are inconsistent at penalizing them for nonhuman-like erroneous output. We also hypothesize that the complexity of a visual scene should be reflected in the linguistic variety of the captions and, in testing this hypothesis, we find that human-generated captions have a dramatically greater degree of lexical, syntactic, and semantic variation. These results have important implications for the design of performance metrics, gauging what deep learning captioning systems really understand in images, and the importance of the task of image captioning for cognitive systems research.

While a majority of children enjoy digital technologies in their daily lives, fewer have the opportunity to learn about code: the sets of computer instructions that control how their devices function. When children and young people learn to code, this can lead to an appreciation of how coding helps them achieve the highest levels of personalisation and control over their media devices. Such engagement then becomes more dynamic and interactive. As a result, coding ability enhances children's experiences with digital devices, helping make young people 'power' users of the web and social media platforms.

Although cyberbullying has attracted a good deal of research attention in recent years, we know much less about how cyberbullying differs by race and ethnicity, and the few studies that have examined this topic have yielded conflicting findings. In the current study, 352 White and Black respondents answered questions about their victimization and perpetration experiences with both traditional bullying and cyberbullying. More than 80% of both White and Black participants indicated that they had been victims of traditional bullying, and over 45% had been victims of cyberbullying at least once. Fewer reported perpetrating either type of bullying. Texting and social media were the most common forms of technology used by both White and Black respondents as well as the most common venues by which cyberbullying occurred. For all participants, cyberbullying victimization was associated with greater suicidal ideation, and traditional bullying victimization was related to higher loneliness, depression, and suicidal ideation. At higher frequencies of traditional bullying victimization, Black respondents reported higher loneliness than did White participants. Understanding patterns of technology use and experiences with cyberbullying victimization and perpetration will help to inform effective strategies for prevention and intervention. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

The ability to recognize that pairs or sets of language expressions “mean the same thing” is a cognitive task for which meaning representation is clearly a central issue. This paper uses the task of paraphrasing to study meaning representation in a cognitive system. The main claim of this paper is that a consequential part of the meaning representation for a natural language expression is a set of language-free structures and processes that are not part of the expression in question. To support our claims, we construct a corpus of paraphrase pairs using a system that has a non-linguistic meaning representation decoupled from the linguistic system that generates natural language from it. This corpus of paraphrase pairs is special in that it represents a full range of syntactic and lexical difference in their constituent sentences. We conduct an extensive analysis comparing the performance of a state-of-the-art neural network model against humans performing the paraphrase detection task. We find that the model deviates significantly from human classification performance, particularly on sentence pairs that conveyed the same meaning while exhibiting significant differences lexically and syntactically. As the neural network model is trained only on linguistic items, the discrepancy points to the existence and necessity of a significant non-linguistic part of meaning formation.

This paper proposes work that applies insights from meaning representation systems for in-depth natural language understanding to representations for self-supervised learning systems, which show promise in developing complex, deeply-nested symbolic structures through self-motivated exploration of their environments. The core of the representation system transforms language inputs into language-free structures that are complex combinations of conceptual primitives, forming a substrate for human-like understanding and common-sense reasoning. We focus on decomposing representations of expectation, intention, planning, and decision-making which are essential to a self-motivated learner. These meaning representations may enhance learning by enabling a rich array of mappings between new experiences and structures stored in short-term and long-term memory. We also argue that learning can be further enhanced when language interaction itself is an integral part of the environment in which the self-supervised learning agent is embedded.

Although social media platforms like Facebook, Instagram, Snapchat, and Twitter are popular platforms for connecting youth, they have increasingly become the locus of harmful social behaviours such as cyberbullying, which can result in disruptive conflicts within schools. Restorative justice seeks to bring together stakeholders to resolve conflicts and build relationships rather than respond to student misbehaviour through punitive approaches. However, restorative justice practitioners and school authorities are grappling with how to address school conflict that arises online. This conceptual paper proposes virtualizing restorative justice practices as a means of engaging more effectively with youth aggression online as well as provide an additional tool to restorative justice practitioners given the time and resource constraints they operate under. By incorporating elements of restorative justice values and principles in responding to online aggression, this proposed online solution aims to mediate and resolve conflict between users, and empower youth participation all with the advantages of connectivity and persistence of social media platforms. We propose a real-time, innovative solution to an issue that is rampantly growing, furthermore bridging the digital divide between adults and students and forming the building blocks for future pilot testing of such an online solution for school districts nationally.

The AAAI 2019 Spring Symposium Series was held Monday through Wednesday, March 25–27, 2019, on the campus of Stanford University, adjacent to Palo Alto, California. The titles of the nine symposia were Artificial Intelligence, Autonomous Machines, and Human Awareness: User Interventions, Intuition and Mutually Constructed Context; Beyond Curve Fitting — Causation, Counterfactuals and Imagination‐Based AI; Combining Machine Learning with Knowledge Engineering; Interpretable AI for Well‐Being: Understanding Cognitive Bias and Social Embeddedness; Privacy‐Enhancing Artificial Intelligence and Language Technologies; Story‐Enabled Intelligence; Toward Artificial Intelligence for Collaborative Open Science; Toward Conscious AI Systems; and Verification of Neural Networks.

There is a dearth of research investigating youths’ experience of grief and mourning after the death of close friends or family. Even less research has explored the question of how youth use social media sites to engage in the grieving process. This study employs qualitative analysis and natural language processing to examine tweets that follow 2 deaths. First, we conducted a close textual read on a sample of tweets by Gakirah Barnes, a gang-involved teenaged girl in Chicago, and members of her Twitter network, over a 19-day period in 2014 during which 2 significant deaths occurred: that of Raason “Lil B” Shaw and Gakirah’s own death. We leverage the grief literature to understand the way Gakirah and her peers express thoughts, feelings, and behaviors at the time of these deaths. We also present and explain the rich and complex style of online communication among gang-involved youth, one that has been overlooked in prior research. Next, we overview the natural language processing output for expressions of loss and grief in our data set based on qualitative findings and present an error analysis on its output for grief. We conclude with a call for interdisciplinary research that analyzes online and offline behaviors to help understand physical and emotional violence and other problematic behaviors prevalent among marginalized communities.

While recent research has shown that “classical” automated planning systems are effective tools for story generation, the success of automated story understanding systems may require integration between commonsense reasoning and more sophisticated forms of planning to make inferences and deductions about the plans and goals of story actors. Methods that decompose abstractions (i.e., tasks or language expressions) into primitives have played an important role for both automated planning systems and automated story understanding systems, but the two areas have remained largely isolated from each other with few overlaps. We argue that this little-explored connection can benefit both areas of research, and this position paper explores the connections between these systems through the common use of primitive decomposition and its variants. Specifically, we present a prototype of a Hierarchical Task Network planner that decomposes natural language input into primitive structures of Conceptual Dependency, a meaning representation designed for in-depth story understanding. We discuss the important challenges, implications, and applications enabled by the establishment of this unique, direct link between planning and story understanding systems.

Although there have been many key advancements in connecting text and perception, computergenerated image captions still lack common sense. As a first step towards constraining these perception mechanisms to commonsense judgment, we have developed reasonableness monitors: a wrapper interface that can explain if the descriptive output of an opaque deep neural network is plausible. These monitor a standalone system that uses careful dependency tracking, commonsense knowledge, and conceptual primitives to explain a perceived scene description to be reasonable or not. If such an explanation cannot be made, it is evidence that something unreasonable has been perceived. The development of reasonableness monitors is work towards generalizing that vision, with the intention of developing a system-construction methodology that enhances robustness at run time by dynamic checking and explaining of the behaviors of scene understanders for reasonableness in context.

Conceptual Dependency (CD) primitives and Image Schemas (IS) share a common goal of grounding symbols of natural language in a representation that allows for automated semantic interpretation. Both seek to establish a connection between high-level conceptualizations in natural language and abstract cognitive building blocks. Some previous approaches have established a CD-IS correspondence. In this paper, we build on this correspondence in order to apply a logic designed for image schemas to selected CD primitives with the goal of formally taking account of the CD inventory. The logic draws from Region Connection Calculus (RCC-8), Qualitative Trajectory Calculus (QTC), Cardinal Directions and Linear Temporal Logic (LTL). One of the primary premises of CD is a minimalist approach to its inventory of primitives, that is, it seeks to express natural language contents in an abstract manner with as few primitives as possible. In a formal analysis of physical primitives of CD we found a potential reduction since some primitives can be expressed as special cases of others.

Computer users today are demanding greater performance from systems that understand and respond intelligently to human language as input. In the past, researchers proposed and built conceptual analysis systems that attempted to understand language in depth by decomposing a text into structures representing complex combinations of primitive acts, events, and state changes in the world the way people conceive them. However, these systems have traditionally been time-consuming and costly to build and maintain by hand. This paper presents two studies of crowdsourcing a parallel corpus to build conceptual analysis systems through machine learning. In the first study, we found that crowdworkers can view simple English sentences built around specific action words, and build conceptual structures that represent decompositions of the meaning of that action word into simple and complex combinations of conceptual primitives. The conceptual structures created by crowdworkers largely agree with a set of gold standard conceptual structures built by experts, but are often missing parts of the gold standard conceptualization. In the second study, we developed and tested a novel method for improving the corpus through a subsequent round of crowdsourcing; In this “refinement” step, we presented only conceptual structures to a second set of crowdworkers, and found that when crowdworkers could identify the action word in the original sentence based only on the conceptual structure, the conceptual structure was a stronger match to the gold standard structure for that sentence. We also calculated a statistically significant correlation between the number of crowdworkers who identified the original action word for a conceptual structure, and the degree of matching between the conceptual structure and a gold standard conceptual structure. This indicates that crowdsourcing may be used not only to generate the conceptual structures, but also to select only those of the highest quality for a parallel corpus linking them to natural language.