CIKM Workshop on Knowledge Injection in Neural Networks
Workshop Schedule (8.00 am PST to 12 noon PST Nov 1st 2021)
Workshop Opening Remarks – 8.00 am PST
Keynote Talk 1 by Subbaro Kambhampati – 8.05 am PST
Invited paper talk: Facts As Experts (Google Research) – 8.35 am PST
Panel discussion – 9.05 am (Panelists: Gary Marcus, Luis Lamb, Vered Shwartz, Partha Talukdar; Moderator: Gadi Singer)
Break – 10.05 – 10.15 am PST
Keynote Talk 2 by Antoine Bosselut 10.15 am to 10.45 am PST
Paper presentations -> 10.45 am to 12.00 noon
Concluding remarks – 12 noon to 12.05 am PST
List of Accepted Papers
Lars Hoffmann, Christian Bartelt and Heiner Stuckenschmidt. Knowledge Injection via ML-based Initialization of Neural Networks
Dhanasekar Sundararaman, Vivek Subramanian, Guoyin Wang, Shijing Si, Dinghan Shen, Dong Wang and Lawrence Carin. Syntactic Knowledge-Infused Transformer and BERT models
Ruben Branco, António Branco, João Silva and João Rodrigues. Commonsense Reasoning: how do Neuro-only and hybrid Neuro-Symbolic approaches compare?
Manisha Verma, Kapil Thadani and Shaunak Mishra. COVID community question answering in presence of external curated knowledge
Hiba Ahsan and Sudarshan Lamkhede. Improving Search Results Ranking Using a Knowledge Graph
Overview
Deep learning has made rapid progress in last decade, with neural network-based language and vision models achieving state of the art performance in various tasks. Yet, purely data-driven neural network models exhibit several issues impacting real world deployment of such models adversely. These include reliance on large quantities of training data, poor robustness, lack of generalization, poor explainability and glaring gaps in implicit and commonsense knowledge.
The availability of rich structured knowledge sources has spurred the research community into exploring Knowledge Injection in Neural Networks (KINN) as a means of mitigating the above-mentioned challenges. This has led to the development of hybrid AI systems which combine the purely data-driven learning of the neural network models with an infusion of knowledge from external sources. Such KINN systems include the development of retrieval augmented neural models, neuro symbolic systems and a plethora of combinations of NNs and knowledge graphs and structured knowledge bases.
Lastly, building AI systems that share human knowledge may have other interesting consequences, such as improved predictability and human-AI trust. But, humans often rely on both explicitly codified and implicitly understood forms of knowledge, and current AI systems seem to lack such knowledge. Hence, we are also interested in systems and methods that explore how knowledge injection can help compensate for the lack of such knowledge in AI systems.
Given the considerable potential and promise of knowledge injection in overcoming the current challenges associated with purely data-driven NN models, we propose a workshop on the theme of Knowledge Injection in Neural Networks (KINN).
Confirmed Speakers/Panelists
Important dates (Tentative)
Paper submission due: July 31st 2021
Notification of acceptance: August 29th 2021
Camera-ready copy due: Sep 4th 2021
Workshop date: November 1st 2021
All deadlines are 23:59 Hours PST
Workshop Organizers/Co-chairs
Vasudev Lal, Intel Labs USA
Yezhou Yang, ASU
Pasquale Minervini, UCL
Somak Aditya, MSR India
Sandya Mannarswamy, Intel India