Papers by Srinath Srinivasa
Many Worlds on a Frame: Characterizing Online Social Cognition
: The theme of the Web Observatory at IIIT Bangalore is "online social cognition." Our ... more : The theme of the Web Observatory at IIIT Bangalore is "online social cognition." Our research aims to understand how social media activity molds collective worldview that in turn impacts several areas of human activity, like business, politics or even social harmony. We first categorize the web into three broad regions or realms: called the social, trigger, and inert realms respectively. The social realm forms the participatory areas of the web, where opinions are actively exchanged and molded. Trigger realm refers to elements like news websites or blogs, whose publishing events often trigger activity in the social realm. The inert realm refers to static web content, that gets used as a source of latent knowledge in the social interactions. The social realm itself is modeled as a "marketplace of opinions" -- where different vested interests invest their opinions in order to fetch returns. Opinions that are "compatible" come together to form one or mor...
Proceedings of the 2007 Pacific Asia conference on Intelligence and security informatics
Google, Inc. (search). ...
Resolving the Dilemma of Responsibility in Multi-agent Flow Networks
Lecture Notes in Computer Science, 2023
Big Data Management for Policy Support in Sustainable Development
Lecture Notes in Computer Science, 2022
A Semi-automatic Approach for Generating Video Trailers for Learning Pathways
Springer eBooks, 2022
Automatic Generation of Coherent Learning Pathways for Open Educational Resources
Lecture Notes in Computer Science, 2019
Learners and educators all over the world have been increasingly relying on the internet for educ... more Learners and educators all over the world have been increasingly relying on the internet for education, thus generating and consuming vast amounts of online learning resources. Selecting appropriate learning resources among them and structuring them in a way that maximises comprehension and skill building is a challenging task. In this work, we propose a model to automatically generate learning pathways from available open learning resources, such that the generated pathways are semantically coherent and pedagogically progressive. The proposed model has two components– a Greedy Generator and a Validator based on Support Vector Machine (SVM) and Long Short-Term Memory (LSTM) models respectively. The Greedy Generator chooses the next resource in the learning pathway based on local considerations and the Validator validates the learning pathway as a whole. They work in tandem with each other connected by a feedback loop. Since we work with open educational resources that lack standard meta-data, we also propose methods to generate metrics that compare a pair of learning resources. The learning pathways generated by our model from a corpus of open learning resources show promising results.
Ontology Augmented Data Lake System for Policy Support
Lecture Notes in Computer Science, 2023
Adaptive Agents and Multi-Agents Systems, May 4, 2015
Evolution of cooperation among self-interested agents is revisited in this paper in the context o... more Evolution of cooperation among self-interested agents is revisited in this paper in the context of "globalization" and "localization" and the effects of entrenchment. Entrenchment is found to be of two types-of knowledge and of acquaintance. While entrenched acquaintances are conducive for trust and hence cooperation, entrenched knowledge leads to paucity in novel strategies. Simulation based studies show that disentrenchment in general, and disentrenchment of knowledge in particular, is conducive to the emergence of cooperation.
Lecture Notes in Computer Science, 2009
Complex Systems, Mar 15, 2022
Evolution of cooperation among self-interested agents is revisited in this paper in the context o... more Evolution of cooperation among self-interested agents is revisited in this paper in the context of globalization and localization. A globalized society is characterized by disentrenchment-or routine interactions between strangers across subcultures. Such interactions are rich in novelty, but also have high levels of distrust and insecurity. A localized society is comprised of clusters of subcultures where most social interactions happen. Each tightly knit subculture is rich in mutual familiarity and trust, but not conducive to the spread of novel ideas. A second dimension is that of utilitarian knowledge. Historically, social acquaintances were the primary (if not the only) source of utilitarian knowledge. With technologies like the internet, diffusion of utilitarian knowledge in a society is no longer modulated by acquaintance networks. This leads us to two different forms of (dis)entrenchment: (dis)entrenchment of knowledge and (dis)entrenchment of acquaintance, leading to four societal configurations. This paper asks how each of the configurations fares with respect to the evolution of cooperation. Entrenchment is represented using well-known network models from the literature, and evolution of cooperation is modeled by the evolutionary version of the iterated prisoners' dilemma game. Based on simulation runs, we note that acquaintance and knowledge are characteristically different aspects. We find that disentrenched knowledge is more conducive for evolution of cooperation in networks rather than disentrenched acquaintances.
An Interactive Simulator for COVID-19 Trend Analysis
Spread of COVID-19 has had a devastating impact in almost all countries with 24 million confirmed... more Spread of COVID-19 has had a devastating impact in almost all countries with 24 million confirmed cases and 0.8 million deaths (as of 30 August 2020)1. This in turn has lead to a catastrophic damage to the global economy. Numerous efforts are in progress across the world to control the pandemic. Multiple strategies to control and limit the spread of this virus are being tried out by Government agencies of various countries. Trend analysis of the infection spread can provide critical inputs to develop intervention measures. To analyse COVID-19 spread, we present an interactive simulator that implements an extension of well known SEIR model for epidemic spread, integrated with network diffusion models and different kinds of community interaction models. The simulator is currently implemented over datasets from two regions in India: Bangalore city and Karnataka state. The simulator provides user control of varying multiple data, region and model specific parameters. Results are presented in the form of interactive heatmaps highlighting the spatial distribution, and line plots showing temporal evolution of the disease. Hence, the simulator captures the spatio-temporal spread of COVID-19, providing a comprehensive picture of the pandemic in a region over a selected time period.
A Systematic Review of Online Learning Platforms for Computer Science Courses
arXiv (Cornell University), Jan 10, 2023
In this paper, we propose an AI based approach to Trailer Generation in the form of short videos ... more In this paper, we propose an AI based approach to Trailer Generation in the form of short videos for online educational courses. Trailers give an overview of the course to the learners and help them make an informed choice about the courses they want to learn. It also helps to generate curiosity and interest among the learners and encourages them to pursue a course. While it is possible to manually generate the trailers, it requires extensive human efforts and skills over a broad spectrum of design, span selection, video editing, domain knowledge, etc., thus making it time-consuming and expensive, especially in an academic setting. The framework we propose in this work is a template based method for video trailer generation, where most of the textual content of the trailer is auto-generated and the trailer video is automatically generated, by leveraging Machine Learning and Natural Language Processing techniques. The proposed trailer is in the form of a timeline consisting of various fragments created by selecting, paraphrasing or generating content using various proposed techniques. The fragments are further enhanced by adding voice-over text, subtitles, animations, etc., to create a holistic experience. Finally, we perform user evaluation with 63 human evaluators for evaluating the trailers generated by our system and the results obtained were encouraging.
Creating Navigable Competency Maps from Learning Resource Corpora
This paper addresses the problem of organising large collections of learning resources in a way t... more This paper addresses the problem of organising large collections of learning resources in a way that is conducive to learners to get a semantic overview of the subject area. We propose mechanisms for automatic creation of a “competency map” that organises a learning space in terms of basic units of learning, each of which is called a competency. The competency map is organised as a 2-dimensional progression space. A progression space not only has a concept of distance between any pairs of competencies, but also a partial ordering that indicates progress made by a learner on reaching some competency. In addition to the idea of progression, a competency map needs to be organised such that, learning resources mapped onto the space can create coherent sequences of learning pathways that can be traversed on the space. To address such requirements, we introduce the concept of topic and resource “volumes” which are used as an indicator of the position of any competency in the progression space. Competency maps are created over several learning corpora, containing hand-curated learning pathways. These pathways were then used as ground truth to evaluate the soundness of the organisation of competencies.
Narrative Plot Comparison Based on a Bag-of-actors Document Model
Comparing documents based on their semantic plot structure or narrative is an important problem i... more Comparing documents based on their semantic plot structure or narrative is an important problem in several application areas. Approaches based on information retrieval methods, latent semantic indexing, sentence embedding, and topic modeling are inadequate to capture the structural elements of the narrative. In this work, we present an abstract "bag-of-actors" document model, meant for comparing, indexing and retrieving documents based on their narrative structures. This model is based on resolving the main entities or actors in the plot, and the corresponding actions associated with them. We use this to compare movie plot summaries from IMDB (Internet Movie Database) to identify movie plots that are remakes or were inspired by one another. Evaluation over a wide range of movie plots from different genres, show encouraging results.
Personal and sensitive information about individuals, often needs to be legitimately exchanged am... more Personal and sensitive information about individuals, often needs to be legitimately exchanged among different stakeholders, to provide services, maintain public health, law and order, and so on. While such exchanges are necessary, they also impose enormous privacy and security challenges. Data protection laws like GDPR specify conditions and the legal capacity in which personal information can be solicited and disseminated further. But there is a dearth of formalisms for specifying legal capacities and jurisdictional boundaries, so that open-ended exchange of sensitive data can be implemented. This paper proposes an extensible framework called Multiverse in which sensitive data can flow across a network through "role tunnels" established based on corresponding legal capacities.
Proceedings of the Second International Conference on Big Data Analytics - Volume 8302
International Conference on Big Data, Dec 16, 2013
Design of the Cogno Web Observatory for Characterizing Online Social Cognition
Lecture Notes in Computer Science, 2018
It is important to occasionally remember that the World Wide Web (WWW) is the largest information... more It is important to occasionally remember that the World Wide Web (WWW) is the largest information network the world has ever seen. Just about every sphere of human activity has been altered in some way, due to the web. Our understanding of the web has been evolving over the past few decades ever since it was born. In its early days, the web was seen just as an unstructured hypertext document collection. However, over time, we have come to model the web as a global, participatory, socio-cognitive space. One of the consequences of modeling the web as a space rather than as a tool, is the emergence of the concept of Web observatories. These are application programs that are meant to observe and curate data about online phenomena. This paper details the design of a Web observatory called Cogno, that is meant to observe online social cognition. Social cognition refers to the way social discourses lead to the formation of collective worldviews. As part of the design of Cogno, we also propose a computational model for characterizing social cognition. Social media is modeled as a “marketplace of opinions” where different opinions come together to form “narratives” that not only drive the discourse, but may also bring some form of returns to the opinion holders. The problem of characterizing social cognition is defined as breaking down a social discourse into its constituent narratives, and for each narrative, its key opinions, and the key people driving the narrative.
Computing Exposition Coherence Across Learning Resources
Lecture Notes in Computer Science, 2018
With increasing numbers of open learning resources on the web that are created and published inde... more With increasing numbers of open learning resources on the web that are created and published independently by different sources, stringing together coherent learning pathways is a challenging task. Coherence in this context means the semantic “smoothness” of transition from one learning resource to the next, i.e., the change in topic distribution and exposition styles between consecutive resources is minimal, and the overall sequence of resources together provides a good learning experience. Towards this end, we present a model to compute exposition coherence between a pair of learning resources, based on representing exposition styles in the form of a random walk. It is based on an underlying hypothesis about exposition styles modelled as a sequence of topical entailments. Evaluation of the presented model on the dataset of learning pathways curated by the teachers of the educational platform Gooru.org, show promising results.
When Extrinsic Payoffs Meet Intrinsic Expectations
Lecture Notes in Computer Science, 2023
Uploads
Papers by Srinath Srinivasa