Events and News

Speaker: Dr. Smita Sirker

Abstract In this talk, I address the domain of epistemic decision making in social contexts. How do epistemic agents form beliefs about something (in the world) given evidential inputs from other people? How do epistemic agents respond and alter their doxastic attitude in the light of testimony (evidence from others)? There have been debates whether testimony generates or transmits knowledge. Jennifer Lackey argues that depending on the reliability of the testifier (as a speaker) testimony can be a source of knowledge. It is interesting to note here that the testifier as a speaker has the potential to be a source of (testimonial) knowledge and not as a believer. Pushing this point a bit more, I will primarily argue how the nature of exchange between the speaker and the hearer creates many dimensions of understanding in an epistemic agent (who is again the hearer in this context), when one considers the presence of evidentials. Thus, a key aspect to look into is the concept of evidentiality (in languages). We focus on the epistemic agent (as the hearer) and the potential ways in which the hearer may be influenced by the nature of exchange with the speaker (the testifier, another epistemic agent) and the use of evidentiality by the speaker. In brief, we explore the factor of evidentiality in understanding testimonies as a source of knowledge (or in a weaker sense, testimonial beliefs).

Speaker: Dr. Vinoo Alluri

Abstract: In this era of big data, there is an unparalleled opportunity to tap into people's natural musical experiences. This occurs across various platforms: those that facilitate individuals in openly expressing themselves and sharing music, such as Twitter, and those that create a secure space for sharing more personal sentiments and connections with music anonymously, like Reddit. Modern computational tools now permit us to extract and connect this data to musical preferences, allowing the extraction of numerous musical aspects, including lyrics. Nevertheless, each platform comes with its own limitations regarding the types of questions that can be asked and potential biases that might exist. In my talk, I will describe studies that have focused on identifying patterns in music listening that correspond to emotional well-being on platforms such as Reddit, Last.fm, and Spotify. The discussion will also encompass how music was used as a coping mechanism during the pandemic and within social movements on Twitter. The presentation will also discuss hurdles and inherent biases within utilizing big data for research on music cognition how these obstacles could be mitigated by integrating extensive data analysis with more conventional techniques like surveys, interviews, followed by controlled experiments.

Speaker: Dr. Aneesha Sharma

Abstract: The presentation will introduce Design Cognition as a research discipline, highlighting the diverse aspects of Design that contribute to the cognitive science knowledge base from a Design perspective. Additionally, it will explore how various Design domains utilize cognitive science insights to improve understanding, benefiting both industry and academic research. Bio: Aneesha Sharma currently serves as the Head of the Department of Design at IIT Delhi. Previously, she led the Information Design Discipline at NID. She holds a bachelor's degree in Graphic Arts from J.J. School of Applied Arts, a Master's degree in Communication Design, and a Doctorate from IDC School of Design, IIT Bombay. With over 24 years of experience, Aneesha has worked extensively in both the Design Industry and academia. At the Department of Design, she instructs courses related to UX Design, Communication Design, and Interaction Design. Within the Department of Design, Aneesha established the Experience Lab, focusing on areas such as Design Cognition, Aesthetics, Product Semantics, and Craft cultures.

Speaker: Dr Marieke van Vugt

Mind-wandering is loosely defined as thoughts that are not produced from the current task. Mind-wandering consists of thoughts that are task-unrelated and stimulus-independent.[1][2] This can be in the form of three different subtypes: positive constructive daydreaming, guilty fear of failure, and poor attentional control.[3] In general, a folk explanation of mind-wandering could be described as the experience of thoughts not remaining on a single topic for a long period of time, particularly when people are engaged in an attention-demanding task.[4] One context in which mind-wandering often occurs is driving. This is because driving under optimal conditions becomes an almost automatic activity that can require minimal use of the task positive network,[5] the brain network that is active when one is engaged in an attention-demanding activity. In situations where vigilance is low, people do not remember what happened in the surrounding environment because they are preoccupied with their thoughts. This is known as the decoupling hypothesis.[6] Studies using event-related potentials (ERPs) have quantified the extent that mind-wandering reduces the cortical processing of the external environment. When thoughts are unrelated to the task at hand, the brain processes both task-relevant and unrelated sensory information in a less detailed manner.[7][8][9] Mind-wandering appears to be a stable trait of people and a transient state. Studies have linked performance problems in the laboratory[10] and in daily life.[11] Mind-wandering has been associated with possible car accidents.[12] Mind-wandering is also intimately linked to states of affect. Studies indicate that task-unrelated thoughts are common in people with low or depressed mood.[13][14] Mind-wandering also occurs when a person is intoxicated via the consumption of alcohol.[15] Studies have demonstrated a prospective bias to spontaneous thought because individuals tend to engage in more future than past related thoughts during mind-wandering.[16] The default mode network is thought to be involved in mind-wandering and internally directed thought,[17] although recent work has challenged this assumption

Speaker: Dr. Tanmoy Chakraborty

Abstract: Large Language Models, despite their astounding reasoning abilities, are not faithful problem solvers. While their abilities are strongly correlated with scale, even humongous models like GPT-3.5 or GPT-4 can become inconsistent reasoners. Recent advances in verbose prompting techniques like chain-of-thought try to elicit step-by-step decomposition so that the model can solve a sequence of simpler problems to finally reach the goal. Augmenting external tools like web search or calculators has also been proposed to offload deterministic tasks. However, foundational language models learn neither problem decomposition nor tool-usage. In this talk, I will present potent solutions towards offloading reasoning subtasks in the case of mathematical problem solving: how to teach an auxiliary (and potentially frugal) language model to coordinate with black-box solvers, symbolic or language model-based, to successfully answer mathematical problems. This talk will focus on successfully teaching language models to perform reasoning from non-human feedback and how rewards beyond just the correctness of the final answer are essential for better learning. Bio: Tanmoy Chakraborty is an Associate Professor of Electrical Engineering and an Associate Faculty member of the Yardi School of AI at IIT Delhi. He leads the Laboratory for Computational Social Systems (LCS2), a research group specializing in NLP and Computational Social Science. His current research primarily focuses on empowering frugal language models for improved reasoning, grounding, and prompting and applying them specifically to two applications -- mental health counselling and Cyber-informatics. Tanmoy obtained his PhD in 2015 from IIT Kharagpur as a Google PhD scholar. Subsequently, he worked as a postdoctoral researcher at the University of Maryland, College Park and as a faculty member at IIIT Delhi. Tanmoy has received numerous awards, including the Ramanujan Fellowship, PAKDD Early Career Award, ACL'23 Outstanding Paper Award, IJCAI'23 AI for Good Award, and several faculty awards/gifts from industries like Facebook, Google, LinkedIn, JP Morgan, and Adobe. He has authored a textbook on "Social Network Analysis". More details may be found at www.tanmoychak.com.

Speaker: Dr. Lizzie Bradford

Abstract Social interactions play an important role in daily life, allowing individuals to build social connections and communicate with others. The success of social interactions can have a direct effect on an individuals' mental and physical health; it is therefore important to understand factors that lead to different success rates in social interactions. One factor that underlies social skills is engagement of social cognitive abilities: the ability to consider both one's own and other people's mental states (perspectives, feelings, beliefs, etc). In this talk, I will present recent data that has examined changes in social cognitive abilities, from both the 'self' (processing of changes from one's own perspective) and 'other' perspective (processing of another person's perspective, which may or may not be aligned with one's own), across the lifespan. Using cross-sectional and longitudinal data, we show observable differences in social cognitive capacities (e.g., self/other belief attribution, visual perspective taking) across the lifespan, from 10-90 years old, in both lab-based and real-world scenarios.