Humanities & Social Sciences

89 items

arXiv Sep 27, 2026
The work formalizes audio description (AD) generation as a constrained optimization problem over three coupled decisions—what visual information to describe, when it can be spoken in gaps between dialogue, and how to formulate it to fit the available time—and proposes a hybrid system in which large language models propose and ground visual elements, estimate their narrative salience, and generate compressed realizations, while a mixed-integer linear program jointly selects and schedules descriptions across a scene under temporal constraints; evaluated on REFRAMED, a benchmark for realistic AD of movies, the approach makes better decisions than prompted LLMs about what to describe and when to describe it, establishing a new state of the art on narrative QA and temporally grounded metrics, w
The work formalizes audio description (AD) generation as a constrained optimization problem over three coupled decisions—what visual information to describe, when it can be spoken in gaps between dialogue, and how to formulate it to fit the available time—and proposes a hybrid system in which large language models propose and ground visual elements, estimate their narrative salience, and generate compressed realizations, while a mixed-integer linear program jointly selects and schedules descriptions across a scene under temporal constraints; evaluated on REFRAMED, a benchmark for realistic AD of movies, the approach makes better decisions than prompted LLMs about what to describe and when to describe it, establishing a new state of the art on narrative QA and temporally grounded metrics, w
Reframing audio description as constrained optimization over what, when, and how lets a hybrid LLM-plus-MILP system set new state of the art on REFRAMED's narrative QA and temporal metrics
The work formalizes audio description (AD) generation as a constrained optimization problem over three coupled decisions—what visual information to describe, when it can be spoken in gaps between dialogue, and how to formulate it to fit the available time—and proposes a hybrid system in which large language models propose and ground visual elements, estimate their narrative salience, and generate compressed realizations, while a mixed-integer linear program jointly selects and schedules descriptions across a scene under temporal constraints; evaluated on REFRAMED, a benchmark for realistic AD of movies, the approach makes better decisions than prompted LLMs about what to describe and when to describe it, establishing a new state of the art on narrative QA and temporally grounded metrics, w
The work formalizes audio description (AD) generation as a constrained optimization problem over three coupled decisions—what visual information to describe, when it can be spoken in gaps between dialogue, and how to formulate it to fit the available time—and proposes a hybrid system in which large language models propose and ground visual elements, estimate their narrative salience, and generate compressed realizations, while a mixed-integer linear program jointly selects and schedules descriptions across a scene under temporal constraints; evaluated on REFRAMED, a benchmark for realistic AD of movies, the approach makes better decisions than prompted LLMs about what to describe and when to describe it, establishing a new state of the art on narrative QA and temporally grounded metrics, w
arXiv Sep 27, 2026
This paper presents an in-situ qualitative study of a persistent, proactive AI agent 'teammate' deployed across multiple teams in a large technology company, finding that the boundaries of the human-agent workplace are actively in flux and trigger breakdowns and negotiations across three areas: tacit rules of collaborative human workflows, the relational boundaries of this new non-human actor, and the redistribution of trust and human agency; the authors use these early micro-negotiations as signals to chart a research, design, and organizational agenda that intentionally preserves human agency.
This paper presents an in-situ qualitative study of a persistent, proactive AI agent 'teammate' deployed across multiple teams in a large technology company, finding that the boundaries of the human-agent workplace are actively in flux and trigger breakdowns and negotiations across three areas: tacit rules of collaborative human workflows, the relational boundaries of this new non-human actor, and the redistribution of trust and human agency; the authors use these early micro-negotiations as signals to chart a research, design, and organizational agenda that intentionally preserves human agency.
After deploying a persistent, proactive AI 'teammate' across multiple teams at a large technology company, researchers observed human-agent workplace boundaries being continually negotiated across collaboration rules, the non-human actor's relational boundaries, and the redistribution of trust.
This paper presents an in-situ qualitative study of a persistent, proactive AI agent 'teammate' deployed across multiple teams in a large technology company, finding that the boundaries of the human-agent workplace are actively in flux and trigger breakdowns and negotiations across three areas: tacit rules of collaborative human workflows, the relational boundaries of this new non-human actor, and the redistribution of trust and human agency; the authors use these early micro-negotiations as signals to chart a research, design, and organizational agenda that intentionally preserves human agency.
This paper presents an in-situ qualitative study of a persistent, proactive AI agent 'teammate' deployed across multiple teams in a large technology company, finding that the boundaries of the human-agent workplace are actively in flux and trigger breakdowns and negotiations across three areas: tacit rules of collaborative human workflows, the relational boundaries of this new non-human actor, and the redistribution of trust and human agency; the authors use these early micro-negotiations as signals to chart a research, design, and organizational agenda that intentionally preserves human agency.
arXiv Sep 27, 2026
In professional development workshops, 27 middle school teachers used a teacher-facing chatbot authoring tool, and by analyzing focus-group interviews alongside configuration and interaction logs the study examined how pedagogical intentions are translated into chatbot configurations and reflected in chatbot behavior, finding that teachers envisioned chatbots as instructional scaffolds offering differentiated support, extending access to assistance, and preserving student thinking within teacher-defined boundaries; configuration analysis showed Purpose primarily captured instructional goals and content focus while Rules more often specified pedagogical behavior, guardrails, and learner-specific adaptations; and log-based evaluation showed stronger alignment for responsiveness (88.
In professional development workshops, 27 middle school teachers used a teacher-facing chatbot authoring tool, and by analyzing focus-group interviews alongside configuration and interaction logs the study examined how pedagogical intentions are translated into chatbot configurations and reflected in chatbot behavior, finding that teachers envisioned chatbots as instructional scaffolds offering differentiated support, extending access to assistance, and preserving student thinking within teacher-defined boundaries; configuration analysis showed Purpose primarily captured instructional goals and content focus while Rules more often specified pedagogical behavior, guardrails, and learner-specific adaptations; and log-based evaluation showed stronger alignment for responsiveness (88.
Twenty-seven middle school teachers configured pedagogical intent in a teacher-facing chatbot authoring tool, and log-based evaluation showed 88.9% alignment for responsiveness and 81.5% for persona versus 70.4% for rules and 59.3% for purpose
In professional development workshops, 27 middle school teachers used a teacher-facing chatbot authoring tool, and by analyzing focus-group interviews alongside configuration and interaction logs the study examined how pedagogical intentions are translated into chatbot configurations and reflected in chatbot behavior, finding that teachers envisioned chatbots as instructional scaffolds offering differentiated support, extending access to assistance, and preserving student thinking within teacher-defined boundaries; configuration analysis showed Purpose primarily captured instructional goals and content focus while Rules more often specified pedagogical behavior, guardrails, and learner-specific adaptations; and log-based evaluation showed stronger alignment for responsiveness (88.
In professional development workshops, 27 middle school teachers used a teacher-facing chatbot authoring tool, and by analyzing focus-group interviews alongside configuration and interaction logs the study examined how pedagogical intentions are translated into chatbot configurations and reflected in chatbot behavior, finding that teachers envisioned chatbots as instructional scaffolds offering differentiated support, extending access to assistance, and preserving student thinking within teacher-defined boundaries; configuration analysis showed Purpose primarily captured instructional goals and content focus while Rules more often specified pedagogical behavior, guardrails, and learner-specific adaptations; and log-based evaluation showed stronger alignment for responsiveness (88.
Terence Tao blog RSS Sep 26, 2026
In this guest blog post, probabilist Ivan Corwin argues for a value-based approach to navigating AI's impact on mathematics, proposing that mathematicians produce value in four loci—society, students, community, and individuals—and calling on the mathematical community to clearly articulate and communicate its value to society and funders while recentering teaching and training.
In this guest blog post, probabilist Ivan Corwin argues for a value-based approach to navigating AI's impact on mathematics, proposing that mathematicians produce value in four loci—society, students, community, and individuals—and calling on the mathematical community to clearly articulate and communicate its value to society and funders while recentering teaching and training.
Probabilist Ivan Corwin proposes a value-based approach to mathematics under AI, framing mathematical value across society, students, community, and individuals, and calling on the community to articulate that value to funders
In this guest blog post, probabilist Ivan Corwin argues for a value-based approach to navigating AI's impact on mathematics, proposing that mathematicians produce value in four loci—society, students, community, and individuals—and calling on the mathematical community to clearly articulate and communicate its value to society and funders while recentering teaching and training.
In this guest blog post, probabilist Ivan Corwin argues for a value-based approach to navigating AI's impact on mathematics, proposing that mathematicians produce value in four loci—society, students, community, and individuals—and calling on the mathematical community to clearly articulate and communicate its value to society and funders while recentering teaching and training.
Journal of Applied Business and Economics Sep 26, 2026
Using a structured closed-ended questionnaire with financial professionals across diverse U.S. lending institutions and analyzing the data through the Technology-Organization-Environment (TOE) framework and Information Asymmetry Theory, the study finds that AI and BI applications significantly enhance the precision, speed, and objectivity of SME credit risk assessment, improve identification of high-risk borrowers, and reduce subjective biases, while institutional readiness, technological infrastructure, skilled personnel, and regulatory alignment emerge as critical enablers and data fragmentation, capital constraints, and model explainability persist as challenges.
Using a structured closed-ended questionnaire with financial professionals across diverse U.S. lending institutions and analyzing the data through the Technology-Organization-Environment (TOE) framework and Information Asymmetry Theory, the study finds that AI and BI applications significantly enhance the precision, speed, and objectivity of SME credit risk assessment, improve identification of high-risk borrowers, and reduce subjective biases, while institutional readiness, technological infrastructure, skilled personnel, and regulatory alignment emerge as critical enablers and data fragmentation, capital constraints, and model explainability persist as challenges.
A survey of financial professionals at U.S. lending institutions finds that AI and BI applications improve the precision, speed, and objectivity of SME credit risk assessment and better identify high-risk borrowers
Using a structured closed-ended questionnaire with financial professionals across diverse U.S. lending institutions and analyzing the data through the Technology-Organization-Environment (TOE) framework and Information Asymmetry Theory, the study finds that AI and BI applications significantly enhance the precision, speed, and objectivity of SME credit risk assessment, improve identification of high-risk borrowers, and reduce subjective biases, while institutional readiness, technological infrastructure, skilled personnel, and regulatory alignment emerge as critical enablers and data fragmentation, capital constraints, and model explainability persist as challenges.
Using a structured closed-ended questionnaire with financial professionals across diverse U.S. lending institutions and analyzing the data through the Technology-Organization-Environment (TOE) framework and Information Asymmetry Theory, the study finds that AI and BI applications significantly enhance the precision, speed, and objectivity of SME credit risk assessment, improve identification of high-risk borrowers, and reduce subjective biases, while institutional readiness, technological infrastructure, skilled personnel, and regulatory alignment emerge as critical enablers and data fragmentation, capital constraints, and model explainability persist as challenges.
发表出处待核验 Sep 25, 2026
This work presents the first user study assessing the reliability of LLM-as-a-judge for evaluating conversational recommendation system (CRS) responses: 20 multi-turn sessions were sampled from TalkPlayData-Challenge, candidate responses were generated by four instruction-tuned LLMs, and 20 music-domain experts produced n=400 ratings on Personalization Quality and Explanation Quality; bootstrapped correlation analysis over 10,000 iterations found moderate positive alignment between LLM judges and human assessments (highest r=0.55 for Gemini-3.1Pro on personalization), outperforming all reference-based baselines.
This work presents the first user study assessing the reliability of LLM-as-a-judge for evaluating conversational recommendation system (CRS) responses: 20 multi-turn sessions were sampled from TalkPlayData-Challenge, candidate responses were generated by four instruction-tuned LLMs, and 20 music-domain experts produced n=400 ratings on Personalization Quality and Explanation Quality; bootstrapped correlation analysis over 10,000 iterations found moderate positive alignment between LLM judges and human assessments (highest r=0.55 for Gemini-3.1Pro on personalization), outperforming all reference-based baselines.
400 ratings from 20 music-domain experts show LLM judges align only moderately with humans (up to r=0.55) yet far exceed BLEU and ROUGE-L reference baselines
This work presents the first user study assessing the reliability of LLM-as-a-judge for evaluating conversational recommendation system (CRS) responses: 20 multi-turn sessions were sampled from TalkPlayData-Challenge, candidate responses were generated by four instruction-tuned LLMs, and 20 music-domain experts produced n=400 ratings on Personalization Quality and Explanation Quality; bootstrapped correlation analysis over 10,000 iterations found moderate positive alignment between LLM judges and human assessments (highest r=0.55 for Gemini-3.1Pro on personalization), outperforming all reference-based baselines.
This work presents the first user study assessing the reliability of LLM-as-a-judge for evaluating conversational recommendation system (CRS) responses: 20 multi-turn sessions were sampled from TalkPlayData-Challenge, candidate responses were generated by four instruction-tuned LLMs, and 20 music-domain experts produced n=400 ratings on Personalization Quality and Explanation Quality; bootstrapped correlation analysis over 10,000 iterations found moderate positive alignment between LLM judges and human assessments (highest r=0.55 for Gemini-3.1Pro on personalization), outperforming all reference-based baselines.
IEEE Spectrum Sep 25, 2026
A 15-person multidisciplinary team from ITESO, Universidad Jesuita de Guadalajara (engineering students, faculty advisors, and IEEE Guadalajara Section volunteers) developed RoboMeshA through the EPICS in IEEE initiative, a portable all-in-one educational platform that brings robotics, computer vision, and AI experiences into high school classrooms lacking robotics laboratories; students connect via a web browser to manually interact with the robot or use its control modes to watch it move and detect and avoid obstacles, the team has built two units and partnered with CETI Colomos and Prepa ITESO high schools to validate the platform in classroom settings, and it is developing a modular coupling framework so that four RoboMeshA robots can operate together.
A 15-person multidisciplinary team from ITESO, Universidad Jesuita de Guadalajara (engineering students, faculty advisors, and IEEE Guadalajara Section volunteers) developed RoboMeshA through the EPICS in IEEE initiative, a portable all-in-one educational platform that brings robotics, computer vision, and AI experiences into high school classrooms lacking robotics laboratories; students connect via a web browser to manually interact with the robot or use its control modes to watch it move and detect and avoid obstacles, the team has built two units and partnered with CETI Colomos and Prepa ITESO high schools to validate the platform in classroom settings, and it is developing a modular coupling framework so that four RoboMeshA robots can operate together.
Mexico ITESO team builds portable robotics teaching platform RoboMeshA under EPICS in IEEE, with two units built and classroom validation at high schools
A 15-person multidisciplinary team from ITESO, Universidad Jesuita de Guadalajara (engineering students, faculty advisors, and IEEE Guadalajara Section volunteers) developed RoboMeshA through the EPICS in IEEE initiative, a portable all-in-one educational platform that brings robotics, computer vision, and AI experiences into high school classrooms lacking robotics laboratories; students connect via a web browser to manually interact with the robot or use its control modes to watch it move and detect and avoid obstacles, the team has built two units and partnered with CETI Colomos and Prepa ITESO high schools to validate the platform in classroom settings, and it is developing a modular coupling framework so that four RoboMeshA robots can operate together.
A 15-person multidisciplinary team from ITESO, Universidad Jesuita de Guadalajara (engineering students, faculty advisors, and IEEE Guadalajara Section volunteers) developed RoboMeshA through the EPICS in IEEE initiative, a portable all-in-one educational platform that brings robotics, computer vision, and AI experiences into high school classrooms lacking robotics laboratories; students connect via a web browser to manually interact with the robot or use its control modes to watch it move and detect and avoid obstacles, the team has built two units and partnered with CETI Colomos and Prepa ITESO high schools to validate the platform in classroom settings, and it is developing a modular coupling framework so that four RoboMeshA robots can operate together.
RESEARCH JOURNAL OF PURE SCIENCE AND TECHNOLOGY Sep 24, 2026
Using a qualitative design based on secondary data and document analysis, this study examines data privacy, algorithmic bias, and decision-making in AI education through the proposed SAFE-T Framework (Stakeholder-Aligned Fairness, Ethics, Transparency in AI-Education), finding persistent gaps in transparency and algorithmic biases that reinforce educational inequities, and arguing for fairness-aware models, participatory policy frameworks, and accountability mechanisms such as fairness audits and regulatory oversight.
Using a qualitative design based on secondary data and document analysis, this study examines data privacy, algorithmic bias, and decision-making in AI education through the proposed SAFE-T Framework (Stakeholder-Aligned Fairness, Ethics, Transparency in AI-Education), finding persistent gaps in transparency and algorithmic biases that reinforce educational inequities, and arguing for fairness-aware models, participatory policy frameworks, and accountability mechanisms such as fairness audits and regulatory oversight.
SAFE-T framework proposes that opaque AI decisions and algorithmic bias in education erode stakeholder trust, calling for fairness audits, explainable AI, and participatory governance
Using a qualitative design based on secondary data and document analysis, this study examines data privacy, algorithmic bias, and decision-making in AI education through the proposed SAFE-T Framework (Stakeholder-Aligned Fairness, Ethics, Transparency in AI-Education), finding persistent gaps in transparency and algorithmic biases that reinforce educational inequities, and arguing for fairness-aware models, participatory policy frameworks, and accountability mechanisms such as fairness audits and regulatory oversight.
Using a qualitative design based on secondary data and document analysis, this study examines data privacy, algorithmic bias, and decision-making in AI education through the proposed SAFE-T Framework (Stakeholder-Aligned Fairness, Ethics, Transparency in AI-Education), finding persistent gaps in transparency and algorithmic biases that reinforce educational inequities, and arguing for fairness-aware models, participatory policy frameworks, and accountability mechanisms such as fairness audits and regulatory oversight.
Proceedings of the ACM on Human-Computer Interaction Sep 23, 2026
In a 2×2 between-subjects laboratory experiment on the “My Dream Team” recommender system manipulating user agency (assignment vs. choice) and heterogeneity criteria (included vs. not included), 332 participants formed 83 four-member teams that completed a 30-minute creativity task; re-ranking recommendations by heterogeneity criteria significantly increased selection of different-race (β=0.69, p<0.01) and different-gender (β=0.68, p<0.01) collaborators, granting agency reinforced homophily and reduced surface-level differences, and nudged teams scored significantly higher on creativity than self-assembled teams (Δ=0.45, p_adj<0.05), with the nudge operating without users' awareness.
In a 2×2 between-subjects laboratory experiment on the “My Dream Team” recommender system manipulating user agency (assignment vs. choice) and heterogeneity criteria (included vs. not included), 332 participants formed 83 four-member teams that completed a 30-minute creativity task; re-ranking recommendations by heterogeneity criteria significantly increased selection of different-race (β=0.69, p<0.01) and different-gender (β=0.68, p<0.01) collaborators, granting agency reinforced homophily and reduced surface-level differences, and nudged teams scored significantly higher on creativity than self-assembled teams (Δ=0.45, p_adj<0.05), with the nudge operating without users' awareness.
Re-ranking recommendations led teams to pick more different-race and different-gender collaborators, and nudged teams outscored self-assembled ones on creativity
In a 2×2 between-subjects laboratory experiment on the “My Dream Team” recommender system manipulating user agency (assignment vs. choice) and heterogeneity criteria (included vs. not included), 332 participants formed 83 four-member teams that completed a 30-minute creativity task; re-ranking recommendations by heterogeneity criteria significantly increased selection of different-race (β=0.69, p<0.01) and different-gender (β=0.68, p<0.01) collaborators, granting agency reinforced homophily and reduced surface-level differences, and nudged teams scored significantly higher on creativity than self-assembled teams (Δ=0.45, p_adj<0.05), with the nudge operating without users' awareness.
In a 2×2 between-subjects laboratory experiment on the “My Dream Team” recommender system manipulating user agency (assignment vs. choice) and heterogeneity criteria (included vs. not included), 332 participants formed 83 four-member teams that completed a 30-minute creativity task; re-ranking recommendations by heterogeneity criteria significantly increased selection of different-race (β=0.69, p<0.01) and different-gender (β=0.68, p<0.01) collaborators, granting agency reinforced homophily and reduced surface-level differences, and nudged teams scored significantly higher on creativity than self-assembled teams (Δ=0.45, p_adj<0.05), with the nudge operating without users' awareness.
Nature News Sep 23, 2026
A Nature news feature reviews educators' and cognitive researchers' concerns that generative AI may erode students' critical thinking, and centres on a Nature Reviews Psychology comment arguing that generative AI can boost learners' performance without promoting the deep cognitive and metacognitive processing required for high-quality learning, while also describing long-running OECD work on teaching and assessing critical thinking.
A Nature news feature reviews educators' and cognitive researchers' concerns that generative AI may erode students' critical thinking, and centres on a Nature Reviews Psychology comment arguing that generative AI can boost learners' performance without promoting the deep cognitive and metacognitive processing required for high-quality learning, while also describing long-running OECD work on teaching and assessing critical thinking.
Critical Thinking in the Age of AI: Separating Performance Gains from Learning
A Nature news feature reviews educators' and cognitive researchers' concerns that generative AI may erode students' critical thinking, and centres on a Nature Reviews Psychology comment arguing that generative AI can boost learners' performance without promoting the deep cognitive and metacognitive processing required for high-quality learning, while also describing long-running OECD work on teaching and assessing critical thinking.
A Nature news feature reviews educators' and cognitive researchers' concerns that generative AI may erode students' critical thinking, and centres on a Nature Reviews Psychology comment arguing that generative AI can boost learners' performance without promoting the deep cognitive and metacognitive processing required for high-quality learning, while also describing long-running OECD work on teaching and assessing critical thinking.

Page 5 · showing 10