Advances in Neuro-Symbolic AI — IBM Seminar Series     

links

Advances in Neuro-Symbolic AI — IBM Seminar Series - overview

Advances in Neuro-Symbolic AI IBM Seminar Series

 

Neuro-symbolic AI combines knowledge-driven symbolic AI and data-driven machine learning approaches. The main goal of this research area is to provide AI with increased capability to derive knowledge and general concepts from data and use them, for example:

• To reason, generalize to other tasks
• Use less data to learn a new task
• Understand causality
• Be more robust
• Be able to move from prediction to intervention
• Support knowledge-based decision making

This series of events will include seminars, tutorials, and panel discussions on topics related to neuro-symbolic AI. Besides scientific and technical advances,  we also envision events devoted to evaluating the impact of such advances on people and society. We also plan to engage with the expert working in various roles in corporate environments to discuss what future AI capabilities are needed in enterprises of various sectors. 

This variety of topics, presentation modalities, and stakeholders will allow the audience of this series to identify the best path to advance AI in a way that is at the same time scientifically inspiring, economically sustainable, and beneficial to society.

Most of the events in this series will be publicly broadcasted live, and the recordings will be made available after the event.

 

---

Upcoming Events

 

5 Apr 2021 (Mon) 3pm-4pm (ET)	Human-AI Interaction: Examples in AI and Finances Manuela M. Veloso (J.P. Morgan AI Research)   In this talk, I will overview opportunities to bring AI to Finances, while focusing on the interaction between AI and users.	Eventbrite for a Calendar invite (Free) Direct link to join the live talk (Webex) Join us on Slack
TBD	Semantic Scholar, NLP, and the Fight Against COVID-19 Oren Etzioni (Allen Institute for AI)   This talk will describe the dramatic creation of the COVID-19 Open Research Dataset (CORD-19) at the Allen Institute for AI and the broad range of efforts, both inside and outside of the Semantic Scholar project, to garner insights into COVID-19 and its treatment based on this data. The talk will highlight the difficult problems facing the emerging field of Scientific Language Processing.	Join live talk
TBD	Title: TBD Michael Witbrock (The University of Auckland)   Abstract: TBD	Join live talk
TBD	Title: TBD Mrinmaya Sachan (ETH Zurich)   Abstract: TBD	Join live talk

 

 

---

Previous Events

 

16 Nov 2020 12pm-1pm (ET)	Neuro-Symbolic AI: Overview and Open Questions Alexander Gray (IBM Research)   Alex will introduce the seminar series by first reviewing various motivations for the emerging topic of neuro-symbolic AI, then posing fundamental questions which we hope that the speakers of the series will each chip away at, including definitions, goals, tasks, methodologies, theory, benchmarks, and tools (with some partial answers so far for each question, to start the discussion).	(1h 10m)
3 Dec 2020 (Thu) 2pm-3pm (ET)	Unifying Logical and Statistical AI with Markov Logic Pedro Domingos (University of Washington)   Intelligent systems must be able to handle the complexity and uncertainty of the real world. Markov logic enables this by unifying first-order logic and probabilistic graphical models into a single representation. Many deep architectures are instances of Markov logic. An extensive suite of learning and inference algorithms for Markov logic has been developed, along with open-source implementations like Alchemy. Markov logic has been applied to natural language understanding, information extraction and integration, robotics, social network analysis, computational biology, and many other areas.	(1h 8m)
10 Feb 2021 (Wed) 3pm-4pm (ET)	Neuralsymbolic AI: A bird's eye view Luís C. Lamb (State of Rio Grande do Sul)   The integration of learning and reasoning has been the subject of growing research interest in AI. However, both areas have been developed under clearly different technical foundations and by separate research communities. Neural-symbolic computing aims at integrating neural learning with symbolic approaches typically used in computational logic and knowledge representation in AI. In this talk, we present an overview of the evolution of neural-symbolic methods, with attention to developments towards integrating machine learning and reasoning into a unified foundation that contributes to explainable AI. We concluded by showing that advances in neural-symbolic computing can lead to the construction of richer AI systems.	(1h 9m)
3 Mar 2021 (Wed) 1pm-2pm (ET)	Predicting Exogenous Effects – Risk and Consequence Michael Haydock (IBM)   Most enterprises that produce forecasts do so solely utilizing the history of the target variable. These time series, upon close inspection, are usually an ensemble of various hidden exogenous variables which are impacting the time series, causing the target variable to take on a distinctive shape. The acquisition and careful monitoring of exogenous data of interest can reveal many potential impacts and possible disruptions to the enterprise, with understanding of how they may impact the forecast of a target variable. Projecting these exogenous data forward to reasonable forecast horizons can provide enough time, so that the enterprise can assess the direction of these risks and event consequences in an effort to react and potentially hedge those risks prior to their emergence. This session will describe such an environment where COVID-19 case data (the event) can be projected forward, assigning risks and volatilities to the U.S. Unemployment Situation and U.S. Per Capita Income (the consequence). An Intelligent Workflow process is proposed and some outcomes illustrated.	(56 m)
5 Apr 2021 (Mon) 3pm-4pm (ET)	Human-AI Interaction: Examples in AI and Finances Manuela M. Veloso (J.P. Morgan AI Research)   In this talk, I will overview opportunities to bring AI to Finances, while focusing on the interaction between AI and users.