JG
IBM at ACS Spring 2025
—
- San Francisco, CA, United States and virtual
This event has ended.
About
The American Chemical Society's spring meeting will convene thousands of chemistry professionals and technologists to discuss the latest trends and advances in the field. This year's event theme is "harnessing the power of data” and IBM will present papers, posters, and demos of how we create technology that will help chemists and material scientists to harness the power of data.
Topics include large scale data ingestion and analysis, leveraging the power of foundation models for prediction, generation and building assistants, to automated chemical synthesis executed in autonomous labs and how these advances are driving research in designing more sustainable materials and in drug discovery.
We invite attendees to visit our booth (#4053) to speak with IBM Researchers and interact with demonstrations of our work.
For presentation times of workshops, demos, and papers see the agenda section below. Note: All times are displayed in your local time.
We look forward to seeing you and telling you more about our latest work and career opportunities at IBM.
Learn more about our work in Quantum computing and Accelerated Discovery.
Career opportunities
Visit us at the IBM booth in the exhibitor area to meet with IBM Researchers to speak about our work and future job opportunities.
- Explore all current IBM Research job openings
- Sign up to be notified of future openings by joining our Talent Network.
Keep up with emerging research and scientific developments from IBM Research. Subscribe to the Future Forward Newsletter.
Speakers
Agenda
- Description:
Speaker: Tim Erdmann Presentation: Oral
- Description:
Speaker: Juliana Jansen Ferreira Presentation: Oral
- Description:
Speaker: Jie Shi Presentation: Oral
- Description:
Speaker: Vidushi Sharma Presentation: Oral
- Description:
Speaker: Tim Erdmann Presentation: Oral
- Description:
Speaker: Nathaniel Park Presentation: Oral
- Description:
Speaker: James L Hedrick Presentation: Oral (Invited Talk) Cooperative Research Award: Symposium in Honor of James L. Hedrick and Yi Yan Yang
- Description:
Speaker: Eduardo Almeida Soares Presentation: Oral
- Description:
Speaker: James L Hedrick Presentation: Oral (Invited Talk) Joe DeSimone 60th Birthday Celebration
- Description:
Speaker: Thiago Reschutzegger Presentation: Oral
- Description:
The IBM Research booth will highlight recent advances in developing technology for accelerated scientific discovery. Interactive demonstrations will give you hands-on experience on:
- A conversational AI assistant for chemical discovery, design and visualization to help chemists navigate complex workflows, give them easy access to a range of cheminformatics tools and databases, and support them visualizing their data
- We will describe IBM’s work on Foundation Models for chemistry and materials, focusing on the 3d atom position modality. We will preview an open-weights model that achieves state of art performance in predicting energies and forces and improves fine-tuning performance for predicting macroscopic properties
- IBM Safer Materials Advisor is an advisor to help users on the identification and substitution of harmful materials and also on the discovery of new materials. This demo focus on the identification of PFAS in products described in SDS (Safe Data Sheet) files.
- The Study Planning Assistant (SPA) is an LLM-based agentic workflow that empowers subject matter experts to break down complex goals into actionable tasks. By integrating four key modules, SPA streamlines the planning, allocation, and execution of tasks, facilitating the discovery of new materials, as demonstrated in a PFAS replacement use case.
- Multi-agent workflows leveraging predictive tools have been demonstrated to assist researchers in iterating identifying and iterating on promising new material designs. In this demo we will demonstrate how leveraging IBM’s Granite LLMs in combination with foundation models for materials can enable effective knowledge synthesis and materials design choices across multiple data modalities.
- A web service to access IBM models for drug discovery, including pretrained and finetuned biomedical foundation models (BMFMs) for molecular tasks including property prediction and molecule generation. A catalogue of tools are available through a web app or APIs, and we demonstrate a user interface that guides a chemist through calling BMFM components to perform molecule prediction, generation, and filtering tasks.
- An agentic workflow prototype that uses an LLM-based planner to solve a task specified in a natural language prompt from a chemist by automatically designing a workflow that calls upon a suite of available tools. We demonstrate discovery of potential alternatives to a patented molecule by calling biomedical foundation models, vectorized molecule property databases, and patent databases.
- Description:
Speaker: Leili Zhang Presentation: Oral
- Description:
Speaker: Tim Erdmann Presentation: Oral
- Description:
Speaker: Victor Shirasuna Presentation: Oral
- Description:
Speaker: Viviane Torres da Silva Presentation: Poster
- Description:
Speaker: Indra Priyadarsini Presentation: Oral
- Description:
Speaker: Leonardo Azevedo Presentation: Oral
- Description:
Speaker: Viviane Torres da Silva Presentation: Oral
- Description:
The IBM Research booth will highlight recent advances in developing technology for accelerated scientific discovery. Interactive demonstrations will give you hands-on experience on:
- A conversational AI assistant for chemical discovery, design and visualization to help chemists navigate complex workflows, give them easy access to a range of cheminformatics tools and databases, and support them visualizing their data
- We will describe IBM’s work on Foundation Models for chemistry and materials, focusing on the 3d atom position modality. We will preview an open-weights model that achieves state of art performance in predicting energies and forces and improves fine-tuning performance for predicting macroscopic properties
- IBM Safer Materials Advisor is an advisor to help users on the identification and substitution of harmful materials and also on the discovery of new materials. This demo focus on the identification of PFAS in products described in SDS (Safe Data Sheet) files.
- The Study Planning Assistant (SPA) is an LLM-based agentic workflow that empowers subject matter experts to break down complex goals into actionable tasks. By integrating four key modules, SPA streamlines the planning, allocation, and execution of tasks, facilitating the discovery of new materials, as demonstrated in a PFAS replacement use case.
- Multi-agent workflows leveraging predictive tools have been demonstrated to assist researchers in iterating identifying and iterating on promising new material designs. In this demo we will demonstrate how leveraging IBM’s Granite LLMs in combination with foundation models for materials can enable effective knowledge synthesis and materials design choices across multiple data modalities.
- A web service to access IBM models for drug discovery, including pretrained and finetuned biomedical foundation models (BMFMs) for molecular tasks including property prediction and molecule generation. A catalogue of tools are available through a web app or APIs, and we demonstrate a user interface that guides a chemist through calling BMFM components to perform molecule prediction, generation, and filtering tasks.
- An agentic workflow prototype that uses an LLM-based planner to solve a task specified in a natural language prompt from a chemist by automatically designing a workflow that calls upon a suite of available tools. We demonstrate discovery of potential alternatives to a patented molecule by calling biomedical foundation models, vectorized molecule property databases, and patent databases.
- Description:
Speaker: Brandi Ransom Presentation: Oral
- Description:
IBM Panelist: Jamie Garcia
Additional Panelists: - Laura Gagliardi (UChicago) - Dr. Nathan Baker (Microsoft) - Moderator: Dr. Todd Krauss
- Description:
Speaker: Tim Erdmann Presentation: Oral
- Description:
This symposium has 10 presentations:
- 2:00 PM-2:45 PM Keynote Presentation (Jay Gambetta)
- 2:45 PM-2:55 PM Intermission
- 2:55 PM-3:40 PM Hands-on Workshop (Pedro Rivero)
- 3:40 PM-3:50 PM Intermission
- 3:50 PM-4:15 PM Lecture: Dr. Mario Motta
- 4:15 PM-4:40 PM Lecture: Dr. Kenneth Merz
- 4:15 PM-4:40 PM Computational chemistry enabled by quantum computers (Kenneth Merz)
- 4:40 PM-4:50 PM Q&A Session
- 4:50 PM-5:00 PM Intermission
- 5:00 PM-6:00 PM Workshop II (Kevin Sung and Pedro Rivero)
- Description:
Speaker: Viviane Torres da Silva Presentation: Poster
- Description:
Speaker: Sara Capponi Presentation: Poster
- Description:
The IBM Research booth will highlight recent advances in developing technology for accelerated scientific discovery. Interactive demonstrations will give you hands-on experience on:
- A conversational AI assistant for chemical discovery, design and visualization to help chemists navigate complex workflows, give them easy access to a range of cheminformatics tools and databases, and support them visualizing their data
- We will describe IBM’s work on Foundation Models for chemistry and materials, focusing on the 3d atom position modality. We will preview an open-weights model that achieves state of art performance in predicting energies and forces and improves fine-tuning performance for predicting macroscopic properties
- IBM Safer Materials Advisor is an advisor to help users on the identification and substitution of harmful materials and also on the discovery of new materials. This demo focus on the identification of PFAS in products described in SDS (Safe Data Sheet) files.
- The Study Planning Assistant (SPA) is an LLM-based agentic workflow that empowers subject matter experts to break down complex goals into actionable tasks. By integrating four key modules, SPA streamlines the planning, allocation, and execution of tasks, facilitating the discovery of new materials, as demonstrated in a PFAS replacement use case.
- Multi-agent workflows leveraging predictive tools have been demonstrated to assist researchers in iterating identifying and iterating on promising new material designs. In this demo we will demonstrate how leveraging IBM’s Granite LLMs in combination with foundation models for materials can enable effective knowledge synthesis and materials design choices across multiple data modalities.
- A web service to access IBM models for drug discovery, including pretrained and finetuned biomedical foundation models (BMFMs) for molecular tasks including property prediction and molecule generation. A catalogue of tools are available through a web app or APIs, and we demonstrate a user interface that guides a chemist through calling BMFM components to perform molecule prediction, generation, and filtering tasks.
- An agentic workflow prototype that uses an LLM-based planner to solve a task specified in a natural language prompt from a chemist by automatically designing a workflow that calls upon a suite of available tools. We demonstrate discovery of potential alternatives to a patented molecule by calling biomedical foundation models, vectorized molecule property databases, and patent databases.
- Description:
Speaker: Priscilla Avegliano Barreira Presentation: Oral
- Description:
Speaker: Eduardo Almeida Soares Presentation: Oral
- Description:
Speaker: Nathaniel Park Presentation: Oral
- Description:
Speaker: Victor Shirasuna Presentation: Oral
Upcoming events
- —
IBM at Interspeech 2025
- | Live
- Rotterdam, Netherlands
- —
Berkeley Innovation Forum 2025 at IBM Research
- San Jose, CA, USA
- —
IBM at Crypto 2025
- | Live
- Santa Barbara, CA, USA and virtual
