https://wiki.aiisc.ai/index.php?title=Special:NewPages&feed=atom&hideredirs=1&limit=50&offset=&namespace=0&username=&tagfilter=Knoesis wiki - New pages [en]2024-03-28T22:42:13ZFrom Knoesis wikiMediaWiki 1.26.2https://wiki.aiisc.ai/index.php?title=Neurosymbolic_Artificial_Intelligence_Research_at_AIISCNeurosymbolic Artificial Intelligence Research at AIISC2024-02-19T16:44:23Z<p>Utkarshanijaimini: /* Journal Publications */</p>
<hr />
<div><br />
= Neurosymbolic AI Overview =<br />
<br />
Humans interact with the environment using a combination of perception transforming sensory inputs from their environment into symbols, and cognition - mapping symbols to knowledge about the environment for supporting abstraction, reasoning by analogy, and long-term planning. Human perception-inspired machine perception, in the context of artificial intelligence (AI), refers to large-scale pattern recognition from raw data using neural networks trained using self-supervised learning objectives such as next-word prediction or object recognition. On the other hand,<br />
machine cognition encompasses more complex computations, such as using knowledge of the environment to guide reasoning, analogy, and long-term planning. Humans can also control and explain their cognitive functions. This seems to require the retention of symbolic mappings from perception outputs to knowledge about their environment. For example, humans can follow and explain the guidelines and safety constraints driving their decision making in safety-critical applications such as health care, criminal justice, and autonomous driving.<br />
<br />
Embodying intelligent behavior in an artificial intelligence system must involve both perception, processing raw data and cognition, using background knowledge to support abstraction, analogy, reasoning, and planning. Symbolic structures represent this background knowledge explicitly. Although neural networks are a powerful tool for processing and extracting patterns from data, they lack explicit representations of background knowledge, hindering the reliable evaluation of their cognition capabilities. Furthermore, applying appropriate safety standards while providing explainable outcomes guided by concepts from background knowledge is crucial for establishing trustworthy models of cognition for decision support.<br />
<br />
==Outcomes Achieved So far==<br />
<br />
#The rapid improvement in language models suggests that they will achieve almost optimal performance levels for large-scale perception. Knowledge graphs are suitable for symbolic structures that bridge the cognition and perception aspects because they support real-world dynamism. Unlike static and brittle symbolic logics, such as first-order logic, they are easy to update. In addition to their suitability for enterprise use cases and established standards for portability, knowledge graphs are part of a mature ecosystem of algorithms that enable highly efficient graph management and querying. This scalability allows for modeling large and complex datasets with millions or billions of nodes.<br />
#We find that combining language models and knowledge graphs are most effective in current implementations. However, it also suggests that future knowledge graphs have the potential to model heterogeneous types of application- and domain-level knowledge beyond schemas. This includes workflows, constraint specifications, and process structures<br />
#Combining such enhanced knowledge graphs with high-capacity neural networks would provide the end user with an extremely high degree of algorithmic- and application-level utility. The concern for safety is behind the recent push to withhold further rollout of generative AI systems such as GPT* as current systems could significantly harm individuals and society without additional guardrails. We believe that guidelines, policy, and regulations can be encoded via extended forms of knowledge graphs, which in turn can provide explainability accountability, rigorous auditing capabilities, and safety. Encouragingly, progress is being made on all these fronts swiftly, and the future looks promising<br />
<br />
==Journal Publications==<br />
# Purohit, H., Shalin, V. L., Sheth, A. P., & Sheth, A. (2020). Knowledge graphs to empower humanity-inspired AI systems. IEEE Internet Computing, 24(4), 48–54. https://doi.org/10.1109/MIC.2020.3013683<br />
# Gaur, M., Faldu, K., & Sheth, A. (2021). Semantics of the black-box: Can knowledge graphs help make deep learning systems more interpretable and explainable? IEEE Internet Computing, 25 (1), Jan-Feb 2021, pp. 51–59. https://doi.org/10.1109/MIC.2020.3031769<br />
# Sheth, A., & Thirunarayan, K. (2021). The duality of data and knowledge across the three waves of AI. IT Professional, 23(3), 35–45. https://doi.org/10.1109/MITP.2021.3070985<br />
#Sheth, A., Gaur, M., Roy, K., & Faldu, K. (2021). Knowledge-intensive language understanding for explainable ai. IEEE Internet Computing, 25(5), 19-24. (Link: https://ieeexplore.ieee.org/abstract/document/9514440)<br />
#Sheth, A., Gaur, M., Roy, K., Venkataraman, R., & Khandelwal, V. (2022). Process knowledge-infused ai: Toward user-level explainability, interpretability, and safety. IEEE Internet Computing, 26(5), 76-84. (Link: https://ieeexplore.ieee.org/abstract/document/9889132)<br />
#Jaimini, U. and Sheth, A., 2022. CausalKG: Causal Knowledge Graph Explainability using interventional and counterfactual reasoning. IEEE Internet Computing, 26(1), pp.43-50. (Link: https://ieeexplore.ieee.org/abstract/document/9706608)<br />
#Sheth, A., Roy, K., & Gaur, M. (2023). Neurosymbolic Artificial Intelligence (Why, What, and How). IEEE Intelligent Systems, 38(3), 56-62. (Lin: https://ieeexplore.ieee.org/abstract/document/10148662/)<br />
#Sheth, A., & Roy, K. (2024). Neurosymbolic Value-Inspired AI (Why, What, and How). IEEE Intelligent Systems. (Link: https://arxiv.org/pdf/2312.09928)<br />
<br />
==Conference, Symposium, and Workshop Publications==<br />
#Pallagani, V., Roy, K., Muppasani, B., Fabiano, F., Loreggia, A., Murugesan, K., ... & Sheth, A. (2024). On the Prospects of Incorporating Large Language Models (LLMs) in Automated Planning and Scheduling (APS). International Conference on Automated Planning and Scheduling (ICAPS). (Link: https://arxiv.org/pdf/2401.02500)<br />
#Zi, Y., Veeramani, H., Roy, K., & Sheth, A. (2024). RDR: the Recap, Deliberate, and Respond Method for Enhanced Language Understanding. AAAI Workshop on Neuro-Symbolic Learning and Reasoning in the era of Large Language Models. (Link: https://openreview.net/forum?id=hNQJI0KS3T)<br />
#Zi, Y., Roy, K., Narayanan, V., & Sheth, A. (2024). Exploring Alternative Approaches to Language Modeling for Learning from Data and Knowledge. AAAI Spring Symposium on Empowering Machine Learning and Large Language Models with Domain and Commonsense Knowledge. (Link: https://scholarcommons.sc.edu/cgi/viewcontent.cgi?article=1619&context=aii_fac_pub)<br />
#Roy, K., Oltramari, A., Zi, Y., Shyalika, C., Narayanan, V., & Sheth, A. (2024). Causal Event Graph-Guided Language-based Spatiotemporal Question Answering. AAAI Spring Symposium on Empowering Machine Learning and Large Language Models with Domain and Commonsense Knowledge. (Link: https://scholarcommons.sc.edu/cgi/viewcontent.cgi?article=1618&context=aii_fac_pub)<br />
#Roy, Kaushik, et al. "Knowledge-Infused Self Attention Transformers." KDD Workshop on Knowledge Augmented Methods for NLP (2023). (Link: https://arxiv.org/pdf/2306.13501)<br />
#Jaimini, Utkarshani, Cory Henson, and Amit Sheth. "An Ontology Design Pattern for Representing Causality." 14th Workshop on Ontology Design Pattern at 22nd International Semantic Web Conference (2023). (Link: https://scholarcommons.sc.edu/cgi/viewcontent.cgi?article=1615&context=aii_fac_pub)<br />
#Roy, Kaushik, et al. "Process Knowledge-infused Learning for Clinician-friendly Explanations." AAAI Second Symposium on Human Partnership with Medical Artificial Intelligence (2023). (Link: https://arxiv.org/pdf/2306.09824)<br />
#Roy, Kaushik, and Vipula Rawte. "TDLR: Top Semantic-Down Syntactic Language Representation." NeurIPS'22 Workshop on All Things Attention: Bridging Different Perspectives on Attention. 2022. (Link: https://attention-learning-workshop.github.io/2022/papers/rawte-tdlr_top_semanticdown_syntactic_language_representation.pdf) <br />
# Keyur Faldu, Amit Sheth, Prashant Kikani, Hemang Akbari. KI-BERT: Infusing Knowledge Context for Better Language and Domain Understanding. arXiv:2104.08145 [cs.CL] https://doi.org/10.48550/arXiv.2104.08145<br />
# Kursuncu, U., Gaur, M., & Sheth, A. (2020). Knowledge infused learning (K-IL): Towards deep incorporation of knowledge in deep learning. Proceedings of the AAAI 2020 Spring Symposium on Combining Machine Learning and Knowledge Engineering in Practice (AAAI-MAKE).<br />
# Gaur, M., Desai, A., Faldu, Keyur, & Sheth, A. (2020). Explainable AI using knowledge graphs. ACM CoDS-COMAD Conference. https://aiisc.ai/xaikg/<br />
# Roy, K., Zhang, Q., Gaur, M., & Sheth, A. (2021). Knowledge infused policy gradients for adaptive pandemic control. Proceedings of the AAAI 2021 Spring Symposium on Combining Machine Learning and Knowledge Engineering (AAAI-MAKE 2021).<br />
# Roy, K., Zhang, Q., Gaur, M., & Sheth, A. (2021). Knowledge infused policy gradients with upper confidence bound for relational bandits, In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, pp. 35-50. Springer, Cham, 2021.<br />
<br />
==Application Paper Publications==<br />
#Wickramarachchi, R., Henson, C., & Sheth, A. (2021). Knowledge-infused learning for entity prediction in driving scenes. Frontiers in big Data, 4, 759110. (Link: https://www.frontiersin.org/articles/10.3389/fdata.2021.759110/full)<br />
#Venkataramanan, R., Padhee, S., Rao, S. R., Kaoshik, R., Rajan, A. S., & Sheth, A. (2023). Ki-Cook: clustering multimodal cooking representations through knowledge-infused learning. Frontiers in Big Data, 6. (Link: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10406211/)<br />
#Wickramarachchi, R., Henson, C., & Sheth, A. (2023, June). CLUE-AD: a context-based method for labeling unobserved entities in autonomous driving data. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 37, No. 13, pp. 16491-16493). (Link: https://ojs.aaai.org/index.php/AAAI/article/view/27089/26861)<br />
#Roy, K., Khandelwal, V., Goswami, R., Dolbir, N., Malekar, J., & Sheth, A. (2023). Demo Alleviate: Demonstrating artificial intelligence-enabled virtual assistance for telehealth: The mental health case. Thirty-Seventh AAAI Conference on Artificial Intelligence. (Link: https://scholarcommons.sc.edu/aii_fac_pub/566/)<br />
#Roy, K., Gaur, M., Rawte, V., Kalyan, A., & Sheth, A. (2023). ProKnow: Process Knowledge for Safety Constrained and Explainable Question Generation for Mental Health Diagnostic Assistance., Frontiers Big Data, 09 January 2023 (Link : https://www.frontiersin.org/articles/10.3389/fdata.2022.1056728/full)<br />
#Roy, K., Khandelwal, V., Surana, H., Vera, V., Sheth, A., & Heckman, H. (2024). GEAR-Up: Generative AI and External Knowledge-based Retrieval Upgrading Scholarly Article Searches for Systematic Reviews. AAAI Conference on Artificial Intelligence. (Link: https://arxiv.org/pdf/2312.09948)<br />
<br />
==Keynotes, Tutorials and Talks==<br />
# Sheth, A., Knowledge-infused NLU for Addiction and Mental Health Research, Keynotes at the MAISoN21, Aug 2021, and the ASONAM21, Sept 2021.[https://www.youtube.com/watch?v=u-06kK9TysA Presentation-GDrive][https://www.youtube.com/watch?v=pRUXTuxm3as Video-MAISoN@IJCAI]<br />
# Sheth, A., Don’t Handicap AI without Explicit Knowledge, Keynotes at the IEEE Services, Sept 2021, and the DEXA2021, Oct 2021. [https://docs.google.com/presentation/d/1ANEZ5_zaB4NSZiroYtfUmFbqUQ9LyE8USu5T_-67hdE/edit#slide=id.ge9fc701584_0_312 Presenation-GDrive][https://www.youtube.com/watch?v=neZKoNDiOQw Video IEEE SErvices][https://www.youtube.com/watch?v=u-06kK9TysA Video-DEXA]<br />
#Amit Sheth, Semantics of the Black-Box: Towards Knowledge-infused Learning --- Keynote talk at the Semantic Machine Learning Workshop in conjunction with the 15th IEEE International Conference on Semantic Computing, February 2021, https://youtu.be/cx-l0XDk9Tw<br />
# Sheth, A., Augmented Personalized Health: an explicit knowledge enhanced neural symbolic Health approach to patient empowerment for managing chronic disease burden. Keynote at the SWAT4HCLS (Semantic Web Applications and Tools for Healthcare and Life Sciences), January 2022. [https://www.slideshare.net/apsheth/augmented-personalized-health-an-explicit-knowledge-enhanced-neurosymbolic-dhealth-approach-to-patient-empowerment-for-managing-chronic-disease-burden Slideshare][https://docs.google.com/presentation/d/1SlUaF1caZqbkYlmhKMuwykybV7k1vzu6/edit#slide=id.p1 Presentation-GDrive][https://www.youtube.com/watch?v=CDg6CMNUdRw Video]<br />
# Amit Sheth, From NLP to NLU: Why we need varied, comprehensive, and stratified knowledge (Neuro-symbolic AI), Keynote at KnowledgeNLP at AAAI2023, February 2023. [http://bit.ly/kNLP2023 Slides] [https://lnkd.in/grzi5UyJ Abstract] [https://youtu.be/xyxQXka6dRY Video]<br />
#Roy, K. (2024). Knowledge-infused Neurosymbolic Artificial Intelligence for Mental Healthcare. Intelligent Clinical Care Center, University of Florida. (Slides: https://lnkd.in/eVCzinAk)<br />
Links to previous years' publications: http://wiki.aiisc.ai/index.php/Advancing_Neuro-symbolic_AI_with_Deep_Knowledge-infused_Learning, and https://wiki.aiisc.ai/index.php?title=EAGER:_Knowledge-guided_neurosymbolic_AI_with_guardrails_for_safe_virtual_health_assistants<br />
<br />
==Funding==<br />
<br />
===EAGER: Advancing Neuro-symbolic AI with Deep Knowledge-infused Learning===<br />
* '''NSF Award#''': 2133842<br />
*'''Award Period of Performance''': 2022-2024<br />
* '''Award Amount:''' $139,999<br />
<br />
===EAGER: Knowledge-guided neurosymbolic AI with guardrails for safe virtual health assistants===<br />
*'''NSF Award #''': 2335967<br />
*'''Award Period of Performance''': 2023-2025<br />
* '''Award Amount:''' $200,000<br />
<br />
==Personnel==<br />
* '''Faculty:''' [http://amit.aiisc.ai/ Amit Sheth] (PI), [https://vigsnar.github.io/ Vignesh Narayanan] (Co-PI)<br />
* '''Graduate Research Assistants:''' [https://www.linkedin.com/in/yuxin-zi/ Yuxin Zi], [https://www.linkedin.com/in/kaushik-roy-b8a323ab/ Kaushik Roy]<br />
* '''AIISC Students:''' [https://www.linkedin.com/in/utkarshanijaimini/ Utkarshani Jamini], [https://www.linkedin.com/in/usha-lokala/ Usha Lokala], [https://www.linkedin.com/in/ruwantw/ Ruwan Wickramarachchi]</div>Adminhttps://wiki.aiisc.ai/index.php?title=Neurosymbolic_Artificial_Intelligence:_Why,_What_and_How%3FNeurosymbolic Artificial Intelligence: Why, What and How?2024-02-10T23:01:47Z<p>Admin: /* Application Papers Publications */</p>
<hr />
<div><br />
= Neurosymbolic AI Overview =<br />
<br />
Humans interact with the environment using a combination of perception - transforming sensory inputs from their environment into symbols, and cognition - mapping symbols to knowledge about the environment for supporting abstraction, reasoning by analogy, and long-term planning. Human perception-inspired machine perception, in the context of AI, refers to large-scale pattern recognition from raw data using foundation models trained with self-supervised learning objectives, such as next-word prediction or object recognition. However, machine cognition encompasses more complex human-like computations, such as using knowledge of the environment to guide reasoning, analogy, and long-term planning. Furthermore, humans can also control and explain their cognitive functions. This requires the retention of symbolic mappings from perception outputs to knowledge about their environment. For example, humans can follow and explain the guidelines and safety constraints that drive their decision-making in safety-critical applications such as healthcare, criminal justice, and autonomous driving. While data-driven neural network-based AI algorithms effectively model machine perception, symbolic knowledge-based AI is better suited for modeling machine cognition. This is because symbolic knowledge structures support explicit representations of mappings from perception outputs to the knowledge, enabling traceability and auditing of the AI system’s decisions. Such audit trails help enforce safety guardrails in applications, such as regulatory compliance and explainability, through tracking the AI system’s inputs, outputs, and intermediate steps. Combining neural networks and knowledge-guided symbolic approaches creates more capable and flexible AI systems. These systems have immense potential to advance both algorithm-level (e.g., abstraction, analogy, reasoning) and application-level (e.g., explainable and safety-constrained decision-making) capabilities of AI systems.<br />
<br />
===Why Neurosymbolic AI?===<br />
Embodying intelligent behavior in an AI system must involve both perception - processing raw data, and cognition - using domain knowledge to support abstraction, analogy, reasoning, and planning. The advantage of symbolic structures is that they represent such domain knowledge explicitly. While neural networks are powerful tools for processing and extracting patterns from data, they lack the explicit representations of domain knowledge required for advanced cognition capabilities. Neurosymbolic AI enables this. Additionally, neurosymbolic AI enables mechanisms for applying appropriate safety standards while providing explainable outcomes guided by concepts from domain knowledge, which is crucial for establishing trustworthy models of cognition for decision support.<br />
<br />
==Methods Papers Publications==<br />
#Roy, K., Oltramari, A., Zi, Y., Shyalika, C., Narayanan, V., & Sheth, A. (2024). Causal event graph-guided language-based spatiotemporal question answering. AAAI Spring Symposium on Empowering Machine Learning and Large Language Models with Domain and Commonsense Knowledge (Link: https://scholarcommons.sc.edu/aii_fac_pub/599/) <br />
#Zi, Y., Roy, K., Narayanan, V., & Sheth, A. (2024). Exploring Alternative Approaches to Language Modeling for Learning from Data and Knowledge. AAAI Spring Symposium on Empowering Machine Learning and Large Language Models with Domain and Commonsense Knowledge. (Link: https://scholarcommons.sc.edu/aii_fac_pub/600/) <br />
#Zi, Y., Veeramani, H., Roy, K., & Sheth, A. (2024). RDR: the Recap, Deliberate, and Respond Method for Enhanced Language Understanding. AAAI Workshop on Neuro-Symbolic Learning and Reasoning in the Era of Large Language Models. (Link: https://openreview.net/forum?id=hNQJI0KS3T)<br />
<br />
==Vision Papers Publications==<br />
#Sheth, A., & Roy, K. (2024). Neurosymbolic Value-Inspired AI (Why, What, and How). IEEE Intelligent Systems. (Link: https://arxiv.org/pdf/2312.09928.pdf)<br />
#Sheth, A., Roy, K., & Gaur. M. (2023). Neurosymbolic AI - Why, What, and How. IEEE Intelligent Systems 38 (3), 56-62. (Link: https://ieeexplore.ieee.org/abstract/document/10148662/)<br />
#Wijesiriwardene, T., Sheth, A., Shalin, V. L., & Das, A. (2023). Why Do We Need Neurosymbolic AI to Model Pragmatic Analogies? IEEE Intelligent Systems, 38(5), 12-16. (Link: https://ieeexplore.ieee.org/document/10269780)<br />
#Gaur, M., & Sheth, A. (2023). Building Trustworthy NeuroSymbolic AI Systems: Consistency, Reliability, Explainability, and Safety. (Link: https://arxiv.org/pdf/2312.06798.pdf)<br />
#Gaur, M., Gunaratna, K., Bhatt, S., & Sheth, A. (2022). Knowledge-infused Learning: A Sweet Spot in Neurosymbolic AI. IEEE Internet Computing, 26(4), 5-11. (Link: https://ieeexplore.ieee.org/document/9841416?denied=)<br />
<br />
==Application Papers Publications==<br />
#Roy, K., Khandelwal, V., Surana, H., Vera, V., Sheth, A., & Heckman, H. (2023). GEAR-Up: Generative AI and External Knowledge-based Retrieval Upgrading Scholarly Article Searches for Systematic Reviews. AAAI Conference on Artificial Intelligence 38. (Link: https://arxiv.org/pdf/2312.09948.pdf)<br />
#Roy, K., Khandelwal, V., Goswami, R., Dolbir, N., Malekar, J., & Sheth, A. (2023). Demo alleviate: Demonstrating Artificial Intelligence Enabled Virtual Assistance for Telehealth: The Mental Health Case. AAAI Conference on Artificial Intelligence 37. (Link: https://arxiv.org/pdf/2304.00025)<br />
#Wickramarachchi, R., Henson, C., & Sheth, A. (2023, June). CLUE-AD: a context-based method for labeling unobserved entities in autonomous driving data. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 37, No. 13, pp. 16491-16493). (Link: https://ojs.aaai.org/index.php/AAAI/article/view/27089)<br />
#Venkataramanan, R., Tripathy, A., Foltin, M., Yip, H. Y., Justine, A., & Sheth, A. (2023). Knowledge Graph Empowered Machine Learning Pipelines for Improved Efficiency, Reusability, and Explainability. IEEE Internet Computing, 27(1), 81-88. (Link: https://ieeexplore.ieee.org/abstract/document/10044293)<br />
<br />
More papers related to Neurosymbolic AI by our group can be found at these links: https://shorturl.at/qxU69 , and https://shorturl.at/cftGK<br />
<br />
==Tutorials==<br />
#Roy, K., Lokala, U., Gaur, M., & Sheth, A. P. (2022, October). Tutorial: Neuro-symbolic AI for Mental Healthcare. In Proceedings of the Second International Conference on AI-ML Systems (pp. 1-3). (Link: https://dl.acm.org/doi/abs/10.1145/3564121.3564817)<br />
<br />
===Summary and Learnings So Far===<br />
#We have developed a set of algorithms to enable the development of neurosymbolic AI methods that incorporate declarative knowledge using knowledge graphs and procedural knowledge from domain-specific knowledge sources. <br />
#Our work shows the effectiveness of combining language models and knowledge graphs. Knowledge graphs have the potential to model heterogeneous types of application and domain-level knowledge beyond schemas. This includes workflows, constraint specifications, and process structures, further enhancing the power and usefulness of neurosymbolic architectures. Combining such enhanced knowledge graphs with high-capacity neural networks would provide the end user with an extremely high degree of algorithmic and application-level utility.<br />
#We find that knowledge graphs are suitable for symbolic structures that bridge the cognition and perception aspects because they support real-world dynamism. Unlike static and brittle symbolic logic, such as first-order logic, they are easy to update. In addition to their suitability for enterprise-use cases and established standards for portability, knowledge graphs are part of a mature ecosystem of algorithms that enable highly efficient graph management and querying. This scalability allows for modeling large, complex datasets with millions or billions of nodes.<br />
<br />
==Funding==<br />
<br />
*'''NSF Award #''': 2335967<br />
*'''Award Period of Performance''': Start Date: 07/01/2021 End Date: 09/30/2025<br />
*'''Project Titles''': EAGER: Knowledge-guided neurosymbolic AI with guardrails for safe virtual health assistants, and Advancing Neurosymbolic AI with Deep Knowledge infused Learning<br />
* '''Award Amount:''' $337000</div>Adminhttps://wiki.aiisc.ai/index.php?title=Smart_manufacturingSmart manufacturing2023-09-14T19:50:00Z<p>Chathurangi: /* PUBLICATIONS */</p>
<hr />
<div>The utilization of AI in manufacturing processes has ushered in a new era of data-driven optimization and problem-solving. AI is a pivotal cornerstone, especially in Industry 4.0 and 5.0 applications, yielding advanced technologies to guide a new era of heightened efficiency, productivity, and refined decision-making. In this capacity, AI emerges as an indispensable force driving transformative advancements within the manufacturing domain, where data-driven insights and smart automation catalyze innovation and shape the future of industrial processes. The assessment of SoA approaches in the manufacturing realm helped us identify three concrete use cases for further investigation: (1) Rare event prediction, (2) Anomaly detection using process knowledge workflow, and (3) Explaining the root cause of an event through causal knowledge.<br />
<br />
<br />
<br />
= PROJECTS =<br />
<br />
== '''Project-1:'''Rare Event Prediction ==<br />
There are different types of events that can be observed within autonomous systems. For example, our consideration on this topic led us to explore “rare events” as they are particularly important for smart manufacturing applications. Rare events are occurrences that take place with a significantly lower frequency than more common regular events. In manufacturing, predicting such events is particularly important, as they lead to unplanned downtime, shortening equipment lifespan, and high energy consumption. The occurrence of events is usually considered rare if observed in 5-10% of all instances, while the occurrence of less than 1% is considered extremely rare. The rarity of events is inversely correlated with the maturity of a manufacturing industry. Typically, the rarity of events causes the multivariate data generated within a manufacturing process to be highly imbalanced, which leads to bias in predictive models. To this end, we conducted a thorough survey to investigate prior art along four dimensions: (i) data acquisition methods, (ii) data preprocessing techniques, (iii) algorithmic approaches, and (iv) evaluation/assessment approaches. Figure X below summarizes the methods we looked at within each dimension<br />
<br />
Regarding modeling rare event prediction, we first evaluated the role of data enrichment approaches combined with supervised machine-learning techniques. Predominantly, we explored using three data enrichment approaches; data augmentation, sampling and imputation. Our study is based on open real-world datasets we obtained from public data sources. Figure 1 includes the approaches to learning from rare event data that we exemplified from the comprehensive survey.<br />
<br />
<br />
===Figure 1: Approaches to learning from rare event data===<br />
<HTML><center><br />
<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vRshYAi7gKK7SOGrVeLa0JW7Sln3Ajov-oEGV-oik_tPv4W9G6tXqRnUnGO_E08v0wj4hjGetcJcr58/embed?start=false&loop=false&delayms=3000" frameborder="0" width="960" height="569" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe><br />
<br />
</center></html><br />
<br />
== '''Project-2:'''Process Knowledge Workflow using PDDL ==<br />
We proposed Process Knowledge Workflow (PKW) using Planning Domain Definition Language (PDDL) to detect anomalies in the assembly line (Figure 2&3). Specifically, the PKW uses multimodal data from robot sensors, conveyor belt sensors and camera to detect anomalies. The images from the camera will be processed using state-of-the-art object detection techniques to identify the objects of interest and the interactions among them. The data from the sensors will be utilized to detect the movement of the desired objects. Using this information, PKW will be constructed for the expected process. Then the PKW can serve as an orchestrator to identify anomalies in the sensor values. Based on the sensor values, the PKW will also have the ability to define and name the anomalies. We conducted an extensive survey of the list of sensors in the assembly line from neXt Future Factories lab and studied the type and range of data being generated. Using this, the first version of the PKW ontology will be constructed. A demonstration of PKW can be found in Figure 2.<br />
<br />
===Figure 2: Process Knowledge Workflow (PKW)===<br />
<br />
<HTML><center><br />
<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vStMonSKXOdt_RBfyShabvp0hbmHrQg8Cwdr5NLqM3lomjmps7nKQGCd1oIFfirUIfAHFpR6bwBa8FI/embed?start=false&loop=false&delayms=3000" frameborder="0" width="960" height="569" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe><br />
<br />
</center></html><br />
<br />
===Figure 3: Monitoring Events at Each Step===<br />
<br />
<br />
<HTML><center><br />
<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vR3fmEgLdG0XpmksqYcy-M1AETVSmUEh9MqsFlOfMrKUra4I6xGeYCBtOkzgIKQoq8aNqEeZWp6cM4N/embed?start=false&loop=false&delayms=3000" frameborder="0" width="960" height="256" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe><br />
</center></html><br />
<br />
== '''Project-3:'''Explaining the root cause of an event through causal knowledge ==<br />
Our goal is to utilize the process knowledge workflow to construct a sequential causal Bayesian network (Figure 4) that captures both the input-output sequence of the process and the agent information. Specifically, we aim to create a model that can track the cause-and-effect relationships between the input commands, robot actions, and pipeline outputs in the assembly process. By doing so, we will be able to predict the impact of a given input/output and action combination on the assembly pipeline. If an anomaly occurs, we will use the causal Bayesian network to identify the possible root cause and explain the event.<br />
<br />
<br />
===Figure 4: Sequential Causal Bayesian Network===<br />
<br />
<br />
<br />
<HTML><center><br />
<br />
<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vRkWFJWuQnnkTEioKmTwITxTjYs683ifiZalIw8Bw4lWujCLX-LXuD9pI3y_HXI9CnR3P4EEs26TIRJ/embed?start=false&loop=false&delayms=3000" frameborder="0" width="960" height="569" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe><br />
</center></html><br />
<br />
== '''Project-4:'''AI Empowered Personalized Education ==<br />
<br />
We have surveyed the latest AI technologies for personalized education and have identified synergies with some active projects on using AI for education at AIISC. There is growing use of KGs along with statistical AI (i.e., neural network based) techniques for EduTech solutions using AI for Education. One use of KG is for capturing domain knowledge. For example, the EduTech company Embibe.com has developed comprehensive KG for all subjects and courses it supports. Another use of KG is for creating personalized KG to capture the history of student’s interaction with an AI based education application (e.g., what are all the topics students has learned and how well so that the AI based application can guide the next learning objective and help teachers and the students with personalized assessment). At AIISC we are using AI for education through gaming (e.g., teach strategies for solving Rubrics cube) and support analogy-based pedagogy to improve learning about complex topics with the aid of analogy with more familiar concepts. We identify Dynamic KG to be another interesting application for personalized education. Specifically, the use of a Dynamic KG would allow us to keep the course materials up to date with the latest developments while catering to the emerging demands of students’ learning goals. <br />
We believe our active project on AI supported Analogy based personalized learning has an interesting use-case for effectively explaining difficult manufacturing concepts to students. For example, consider the case where students are learning about a new concept: Data Interoperability. Using the analogy about the role of a translator in multilingual communication, we can explain the need for data interoperability to ensure seamless communication between different computer systems and software applications. This approach was tested in the workforce development context as part of the Summer AI Camp for High School students (https://aiisc.ai/safari/).<br />
<br />
= TEAM MEMBERS =<br />
<br />
'''Advised By -''' [https://en.wikipedia.org/wiki/Amit_Sheth Amit Sheth]<br />
<br />
#[https://www.linkedin.com/in/ruwantw/ Ruwan Wickramarachchi]<br />
#[https://www.linkedin.com/in/chathurangi-shyalika-1b89229b/ Chathurangi Shyalika]<br />
#[https://www.linkedin.com/in/utkarshanijaimini/ Utkarshani Jamini]<br />
#[https://www.linkedin.com/in/revathy-venkataramanan/ Revathy Venkatramanan]<br />
#[https://www.linkedin.com/in/vishalpallagani/ Vishal Pallagani]<br />
<br />
<br />
<br />
= PUBLICATIONS =<br />
<br />
# Shyalika, C., Wickramarachchi, R., & Sheth, A., A Comprehensive Survey on Rare Event Prediction. arXiv preprint "https://arxiv.org/pdf/2309.11356.pdf", 2023<br />
# Wickramarachchi, R., Henson, C., & Sheth, A., CLUE-AD: A Context-Based Method for Labeling Unobserved Entities in Autonomous Driving Data. 37th AAAI Conference on Artificial Intelligence (AAAI), February, 2023. <br />
# Wickramarachchi, R., C. Henson & A. Sheth, "Knowledge-Based Entity Prediction for Improved Machine Perception in Autonomous Systems," in IEEE Intelligent Systems, vol. 37, no. 5, pp. 42-49, 1 Sept.-Oct. 2022, doi: 10.1109/MIS.2022.3181015. <br />
# Kalach, F. E., R. Wickramarachchi, R. Harik & A. Sheth, "A Semantic Web Approach to Fault Tolerant Autonomous Manufacturing," in IEEE Intelligent Systems, vol. 38, no. 1, pp. 69-75, 1 Jan.-Feb. 2023, doi: 10.1109/MIS.2023.3235677. <br />
# Jaimini, Utkarshani, Tongtao Zhang, Georgia Olympia Brikis, & Amit Sheth. "iMetaverseKG: Industrial Metaverse Knowledge Graph to Promote Interoperability in Design and Engineering Applications." IEEE Internet Computing 26, no. 6 (2022): 59-67. <br />
# Jaimini, Utkarshani, & Amit Sheth. "CausalKG: Causal Knowledge Graph Explainability using interventional and counterfactual reasoning." IEEE Internet Computing 26, no. 1 (2022): 43-50. <br />
# R. Venkataramanan, A. Tripathy, M. Foltin, H. Y. Yip, A. Justine & A. Sheth, "Knowledge Graph Empowered Machine Learning Pipelines for Improved Efficiency, Reusability, and Explainability," in IEEE Internet Computing, vol. 27, no. 1, pp. 81-88, 1 Jan.-Feb. 2023, doi: 10.1109/MIC.2022.3228087. <br />
# Shyalika, C., Wickramarachchi, R., & Sheth, A., Towards Rare Event Prediction in Manufacturing Domain. CSE Research Symposium, University of South Carolina, April 14, 2023. [Poster presentation] <br />
# Shyalika, C., Wickramarachchi, R., & Sheth, A., Multivariate Data Augmentation for Rare Event Prediction in Manufacturing. CRA-WP Grad Cohort for Women, April 21, 2023. [Poster presentation] <br />
# Shyalika, C., Wickramarachchi, R., Harik, R., & Sheth, A., Evaluating the Role of Multivariate Data Augmentation Towards Rare Event Prediction in Manufacturing. Discover USC Symposium, University of South Carolina, April 21, 2023. [Poster presentation]<br />
<br />
=Blog Posts=<br />
# Shyalika, C., Wickramarachchi, R., & Sheth, A., A Comprehensive Survey on Rare Event Prediction. AIHub, Available at "https://aihub.org/2023/11/22/a-comprehensive-survey-on-rare-event-prediction/", 2023<br />
<br />
=FUNDING=<br />
<br />
<br />
* '''NSF Award#''': 2119654<br />
* '''RII Track 2 FEC: Enabling Factory to Factory (F2F) Networking for Future Manufacturing'''<br />
* '''Timeline:''' Oct 2021 - Sept 2025<br />
* '''Award Amount:''' $739,239<br />
------------------------------------------------------------------------------------------------<br />
<br />
* '''NSF Award#''': 2133842<br />
* '''EAGER: Advancing Neuro-symbolic AI with Deep Knowledge-infused Learning'''<br />
* '''Timeline:''' 01 July 2021 - 30 June 2022<br />
* '''Award Amount:''' $139,999</div>Chathurangihttps://wiki.aiisc.ai/index.php?title=Computational_Analogy_MakingComputational Analogy Making2023-09-01T14:10:19Z<p>Thiliniiw: /* Publications */</p>
<hr />
<div>== Background and motivation ==<br />
<br />
A human’s ability to identify objects/ situations in one context as similar to objects/ situations in another context is identified as analogy making. We largely follow the steps listed below when we make analogies.<br />
'''Building representations:'''<br />
These could be hierarchical graphical structures of entities and relationships between the entities in domains.<br />
'''Mapping:'''<br />
Finding corresponding elements in each structure is performed by mapping<br />
Usually, mapping is done from a source/ base domain (a familiar domain) to a target domain (a new unfamiliar domain).<br />
'''Inference''':<br />
Once the mapping is complete, knowledge from the source domain is transferred to the target domain. Some relationships and entities may be absent in the target domain, that are present in the base domain. Identifying these and completing them can be identified as inference.<br />
<br />
There have been several attempts at formalizing the process of human analogy making and representing the same computationally through symbolic, connectionist, and hybrid approaches. Gentner’s Structure Mapping Theory(SMT) [https://groups.psych.northwestern.edu/gentner/papers/Gentner83.2b.pdf] is paramount in the domain.<br />
<br />
<center><br />
{{#ev:youtube|https://www.youtube.com/watch?v=hl_dGTSgTN4&list=PLqJzTtkUiq577Rc1HpX4iE1_ntNeuppzA&index=9|500|}}<br />
</center><br />
<br />
== People ==<br />
*'''Artificial Intelligence Institute University of South Carolina'''<br />
**[https://www.linkedin.com/in/thilini-w/ Thilini Wijesiriwardene]<br />
**[https://sc.edu/study/colleges_schools/engineering_and_computing/faculty-staff/amitsheth.php Dr. Amit P. Sheth]<br />
**[https://people.wright.edu/valerie.shalin Dr. Valerie Shalin]<br />
**[http://www.amitavadas.com/ Dr. Amitava Das]<br />
<br />
*'''External Collaborators'''<br />
**[https://www.linkedin.com/in/nitin-jain-48a4b318b/ Dr. Nitin Jain]<br />
<br />
== Publications ==<br />
'''1. Towards efficient scoring of student-generated long-form analogies in STEM[https://ceur-ws.org/Vol-3389/ICCBR_2022_Workshop_paper_106.pdf]'''<br />
<br />
Switching from an analogy pedagogy based on comprehension to analogy pedagogy based on production raises an impractical manual analogy scoring problem. Conventional symbol-matching approaches to computational analogy evaluation focus on positive cases and challenge computational feasibility. This work presents the Discriminative Analogy Features (DAF) pipeline to identify the discriminative features of strong and weak long-form text analogies. We introduce four feature categories (semantic, syntactic, sentiment, and statistical) used with supervised vector-based learning methods to discriminate between strong and weak analogies. Using a modestly sized vector of engineered features with SVM attains a 0.67 macro F1 score. While a semantic feature is the most discriminative, out of the top 15 discriminative features, most are syntactic. Combining these engineered features with an ELMo-generated embedding still improves classification relative to an embedding alone. While an unsupervised K-Means clustering-based approach falls short, similar hints of improvement appear when inputs include the engineered features used in supervised learning.<br />
<br />
[[File:Scoring of student-generated.png|600x600px]]<br />
<br />
<br />
'''2. ANALOGICAL - A Novel Benchmark for Long Text Analogy Evaluation in Large Language Models [https://aclanthology.org/2023.findings-acl.218/]'''<br />
<br />
Over the past decade, analogies, in the form of word-level analogies, have played a significant role as an intrinsic measure of evaluating the quality of word embedding methods such as word2vec. Modern large language models (LLMs), however, are primarily evaluated on extrinsic measures based on benchmarks such as GLUE and SuperGLUE, and there are only a few investigations on whether LLMs can draw analogies between long texts. In this paper, we present ANALOGICAL, a new benchmark to intrinsically evaluate LLMs across a taxonomy of analogies of long text with six levels of complexity – (i) word, (ii) word vs. sentence, (iii) syntactic, (iv) negation, (v) entailment, and (vi) metaphor. Using thirteen datasets and three different distance measures, we evaluate the abilities of eight LLMs in identifying analogical pairs in the semantic vector space. Our evaluation finds that it is increasingly challenging for LLMs to identify analogies when going up the analogy taxonomy.<br />
<br />
[[File:Analogical poster.png|900x900px]]<br />
<br />
<br />
'''3. Why Do We Need Neuro-symbolic AI to Model Pragmatic Analogies? [https://arxiv.org/pdf/2308.01936.pdf Extended Version], [https://www.computer.org/csdl/magazine/ex/2023/05/10269780/1QWO61CnTS8 Abridged Version published in IEEE Intelligent Systems]'''<br />
<br />
A hallmark of intelligence is the ability to use a familiar domain to make inferences about a less familiar domain, known as analogical reasoning. In this article, we delve into the performance of Large Language Models (LLMs) in dealing with progressively complex analogies expressed in unstructured text. We discuss analogies at four distinct levels of complexity: lexical analogies, syntactic analogies, semantic analogies, and pragmatic analogies. As the analogies become more complex, they require increasingly extensive, diverse knowledge beyond the textual content, unlikely to be found in the lexical cooccurrence statistics that power LLMs. To address this, we discuss the necessity of employing Neuro-symbolic AI techniques that combine statistical and symbolic AI, informing the representation of unstructured text to highlight and augment relevant content, provide abstraction and guide the mapping process. Our knowledge-informed approach maintains the efficiency of LLMs while preserving the ability to explain analogies for pedagogical applications.<br />
<br />
[[File:Pragmatic analogies.png|650x650px]]<br />
<br />
'''4. On the Relationship between Sentence Analogy Identification and Sentence Structure Encoding in Large Language Models [https://aclanthology.org/2024.findings-eacl.31.pdf]'''<br />
<br />
The ability of Large Language Models (LLMs) to encode syntactic and semantic structures of language is well examined in NLP. Additionally, analogy identification, in the form of word analogies are extensively studied in the last decade of language modeling literature. In this work we specifically look at how LLMs’ abilities to capture sentence analogies (sentences that convey analogous meaning to each other) vary with LLMs’ abilities to encode syntactic and semantic structures of sentences. Through our analysis, we find that LLMs’ ability to identify sentence analogies is positively correlated with their ability to encode syntactic and semantic structures of sentences. Specifically, we find that the LLMs which capture syntactic structures better, also have higher abilities in identifying sentence analogies.<br />
<br />
[[File:Thilini_poster_EACL.png|900x900px]]<br />
<br />
== Talks ==<br />
1. Prof. Amit Sheth presented three of the several of #AIISC 's projects that involve AI for Education at the "Integrating AI into Higher Education: A Faculty Panel Discussion" today, at the event "Exploring the Intersection of AI Teaching" organized by the Center for Teaching Excellence.<br />
[https://docs.google.com/presentation/d/1Egv5wlqyd2ouZWie4bzADOF9WQot_1O85a3aov8mXxc/edit#slide=id.p]<br />
<br />
2. Session we conducted at the highschool summer camp on learning through analogy <br />
<center><br />
{{#ev:youtube|https://www.youtube.com/watch?v=bPKYjlF8-AM&list=PLqJzTtkUiq54WdiV3FG8uAvdwLsWR-phT&index=10|500|}}<br />
</center></div>Dipeshhttps://wiki.aiisc.ai/index.php?title=Systematic_Review_Assitance_using_Leveraging_Background_Knowledge_and_Language_ModelsSystematic Review Assitance using Leveraging Background Knowledge and Language Models2023-08-28T15:29:36Z<p>Admin: </p>
<hr />
<div>'''A collaborations between USC Libraries and the Artificial Intelligence Institute of South Carolina (AIISC)'''<br />
<br />
=Overview=<br />
<br />
The proliferation of artificial intelligence (AI) technologies, e.g., Microsoft’s recent integration of ChatGPT into its Bing search architecture, showcase AI’s immense potential for shaping information search and discovery for educational purposes. For example, students pursuing higher education at universities often need to review various subjects and topics systematically. For this, students consult expert librarians trained in finding and evaluating the information in university libraries. Explainable AI systems have the potential to assist expert librarians in guiding student users through the systematic review process. Moreover, if developed responsibly and with input from expert humans, such a tool could help scale information literacy interventions whose reach is currently limited by employee availability (e.g., time and bandwidth limitations). We propose building an AI pipeline to assist librarians with structured, systematic review search processes. Our proposed pipeline will process student inputs in natural language and reformulate the inputs as structured queries using structured background knowledge. Furthermore, our system will generate explanations of the query reformulation to an expert human who will be involved in developing the system to enable continuous feedback-based refinements.<br />
<br />
=The Systematic Review Process=<br />
The review process involves the following steps: <br />
# Defining the research question: Formulating a clear, well-defined research question of appropriate scope. Often a student needs help to define a research problem precisely and interacts with the expert librarian for help with this effort. <br />
# Developing a review protocol/criteria: This step is often carried out in parallel with the first step and results in defining the terminology and topics that inform the development of the research question<br />
# Developing inclusion and exclusion criteria: The student needs to understand and determine whether the review will include a particular study. For this, they provide well-defined inclusion-exclusion criteria. <br />
Steps 1, 2, and 3 correspond to the steps; identify the issue and determine the question and Write a plan for the review (protocol) in the Figure 1. The remaining steps involve searching a database and using existing machine learning tools to help with the later stages, including article screening, data extraction, and the risk of bias assessment. In this proposal, we aim to design AI technologies to help with steps 1, 2, and 3.<br />
<br />
=Proposed Methodology=<br />
Our proposed system leverages background knowledge and language models enabled tools such as ChatGPT to perform six steps. We will use the running example in Figure 2. to explain the six steps. <br />
The steps are as follows: <br />
# Seed Concept Identification: We will use natural language processing tools to obtain seed concepts from the student’s query. In the figure example, the seed concept is Hepatitis A.<br />
# Concept and Relations Expansion using Background Knowledge: We will leverage our knowledge graph extraction tool to obtain subgraphs corresponding to the seed concepts in the student’s query. The figure shows subgraphs of concepts connected to the seed concept Hepatitis A and the relationships that connect them. The next two steps are an advanced form of Prompt Engineering assistance.<br />
# Query Expansion using the Relevant Terminology and Topics: From step 2, we obtain the expanded concepts and relations from path traversals on the knowledge graph (rooted in the seed concept). The figure shows the relationships and concepts obtained for the seed concept Hepatitis A as causes, diagnoses, affects, associated with, complicates, and Acetaminophen. <br />
# Query Suggestions to the Student: The concepts and relationships from step 3 are fed into a language model such as ChatGPT with an appropriate prompt to obtain a set of reformulated queries. The reformulated queries are submitted as suggestions to the student. The figure shows the prompt to ChatGPT, and the reformulated queries it generates, e.g., “What are the causes of Hepatitis A and how is it diagnosed?” <br />
# Expert Librarian Review: The generated reformulated queries from 4 are also submitted to the expert librarian for review. Librarians are experts in teaching others to find and evaluate information and regularly work through the query process. As a result, they possess an unparalleled understanding of user behavior and the features and limitations of the systems they work with. Thus, our system presents the librarian with an explanation that consists of user-friendly prompt templates annotated with concepts and relationships from the background knowledge. The librarian can then suggest modifications to them to obtain satisfactory reformulated queries. For example, the figure's prompt template is "Formulate five prompt queries with the keywords: causes, diagnoses, affects, associated with, complicates, Acetaminophen". The librarian's suggestions are used to refine step 2. Background knowledge is necessary for contextual and relevant output, without which ChatGPT might generate irrelevant and incoherent queries. Figure 3. shows examples of explanations generated by ChatGPT without background knowledge (right) vs. using background knowledge (left). ChatGPT’s explanations without a controlled vocabulary obtained using background knowledge can include irrelevant and unspecific information. Including the background knowledge produces relevant and targeted outputs more suited for systematic reviews. The librarian may also analyze the safety of the generated queries. For example, the background knowledge contains information about how acetaminophen could be misused. Our system incorporates controls to avoid showing information on that since college students are especially vulnerable to such information. Similar controls can be placed to extend safety to include the relevant ethics and bias issues. Note that no one can guarantee complete safety as safety can be highly context-sensitive. For example, it may be appropriate for addiction researchers to learn how a drug is abused (e.g., through higher doses, combining multiple drugs, snorting, etc.). However, the same may not be suitable for undergraduate students.<br />
# Structured Query Construction: After finalizing a reviewed query, the language model can generate a structured query in the format supported by the underlying library database. The query can then be submitted to the librarian to review the query specifics (e.g., the mappings to different schema elements). For example, the figure shows a Resource Description Framework (RDF) query for the reformulated query “What are the causes of Hepatitis A, and how is it diagnosed?”. The language model knows schema elements from sources such as DBPedia and Schema.org and uses this information to formulate the structured query.<br />
<br />
https://i.postimg.cc/MT8Zr4qN/USCLibrary-Figure1.png<br />
<br />
=Conclusion and Deliverables=<br />
We propose to develop an AI tool that is explainable (and optionally safer). It inputs student-submitted natural language queries and uses background knowledge to guide a language model for generating reformulated queries. The system will submit the reformulated queries and the explanation (the prompt, concepts, and relationships used to obtain the reformulated queries) to the expert librarian for review. Upon review and finalization of the queries, the language model translates the reformulated natural language queries into structured queries for deployment across major databases. <br />
* Deliverable 1. An AI tool to extract the relevant background knowledge from seed concepts identified in student-submitted natural language queries. We will develop the algorithms and mechanisms for identifying seed concepts and retrieving relevant background knowledge with continuous feedback from expert librarians. <br />
* Deliverable 2. An AI-based query expansion and reformulation module that reads in the seed concepts and the background knowledge and fills out a prompt template to generate reformulated queries. The prompt templates will be designed with expert supervision to obtain optimal results in the least amount of time. The objective is to make the results explainable and, optionally, safer.<br />
* Deliverable 3. An AI tool to translate the reformulated natural language queries into structured queries for deployment across major university databases. For this, we will incorporate knowledge of the database systems and their limitations for literature review in consultation with the expert librarian.<br />
<br />
=Outcomes=<br />
The project has been completed and published (publications details below)<br />
* Roy, K., Khandelwal, V., Surana, H., Vera, V., Sheth, A., & Heckman, H. (2024). GEAR-Up: Generative AI and external knowledge-based retrieval upgrading scholarly article searches for systematic reviews. Proceedings of the 38th AAAI Conference on Artificial Intelligence (Record link: https://scholarcommons.sc.edu/aii_fac_pub/593/).</div>Admin