All partnering requests
Quantum machine learning for predicting molecular spectral properties
  • Background
  • What we're looking for
  • What we can offer you
  • Who we are
  • Q&A
Have questions about this request?
Get them answered by the team at BASF.
Background

Spectral properties of molecules are crucial for understanding and analyzing chemical reactions. These properties result from interactions between molecules and electromagnetic radiation, such as ultraviolet (UV), visible (Vis), and infrared (IR) light. By studying these interactions, scientists can gain valuable insights into the molecular structures, dynamics, reaction environments, and material properties. 

 

Currently, we utilize classical simulation methods, such as density functional theory (DFT), to predict molecular spectral properties. While such classical simulation methods provide reliable results, they have limitations in computational efficiency and accuracy, especially for larger, more complex molecules or reactions. To overcome these challenges and enhance predictive power, BASF is interested in exploring Quantum Machine Learning (QML) techniques, which have the potential to significantly improve both the speed and precision of spectral property predictions.

What we're looking for

We are looking for promising QML methods with the potential to exceed classical methods in terms of speed and accuracy. In a joint research project, we would like to evaluate the proposed QML method.

The developed method should ultimately be applicable to different molecules, perform well on provided datasets, and be demonstrated on similar use cases.

Solutions of interest include:
  • Quantum computing algorithms that can be adapted to predict specific spectral properties of molecules
Our must-have requirements are:
  • Clear, high-level description of the quantum architecture
  • Strong rationale for potential quantum advantage
  • Provide relevant references that support your QML approach
What's out of scope:
  • Solutions that require external proprietary datasets.
  • Black-box approaches - we would like to understand the QML method for joint research.
  • Purely classical approaches – we are interested in quantum computing solutions.
Acceptable technology readiness levels (TRL):
Levels 1-4
What we can offer you
Eligible partnership models:
Sponsored research
Benefits:
Sponsored Research
Funding will be proposal dependent, mutually agreed upon and milestone dependent.
Expertise
Partners will have access to internal experts.
Who we are

As a global chemical company, we operate in diverse markets with varying requirements. To ensure responsible conduct and maintain our license to operate, we are guided by our core values and standards. These values shape our daily actions and foster respectful relationships with colleagues, customers, and partners.

Our teams are committed to putting health and safety first, making sustainability a part of every decision, and meeting strict compliance and environmental standards. We embrace diversity—in people, perspectives, and experiences—and promote open feedback grounded in honesty, respect, and trust. We see setbacks as opportunities to learn and improve.

 Learn more
Reviewers
Q&A with BASF

The Q&A is now closed.

Sort by:
Q.
What would be the duration and approximate budget for the project?
5
A.
This would be subject to further discussions, contract negotiations and project specifics.
NS
Nicole Klein Stocke, Community Manager, BASF
January 22, 2025
Is this response helpful?
0
0
Q.
Hi there, Do you provide the data to train the model?
2
A.
Hi Wendy, thank you for your question! We can provide data to train the model and/or suggest a suitable publicly available dataset depending on the method and project plan.
NS
Nicole Klein Stocke, Community Manager, BASF
January 7, 2025
Is this response helpful?
1
0
Q.
What are the limitations of IP associated to the project? Is publication of novel results/research dissemination possible?
2
A.
In such a project and application, we would aim for a publication. However, details would be subject to contract negotiations.
NS
Nicole Klein Stocke, Community Manager, BASF
January 22, 2025
Is this response helpful?
0
0
Q.
What is the maximum tolerable runtime for the algorithm to run on a single molecule?
2
A.
We are not able to directly answer this as it is a trade off between accuracy and speed.
NS
Nicole Klein Stocke, Community Manager, BASF
January 22, 2025
Is this response helpful?
0
0
Q.
Can you be more precise about which kinds of spectral properties you are interested in?
2
A.
We are interested in optical spectral properties. We would be interested in the molecular low-lying energies labelled by different spin states or/and electron configurations.
NS
Nicole Klein Stocke, Community Manager, BASF
January 22, 2025
Is this response helpful?
0
0
Q.
Are you interesting in predicting averages over conformers, or making predictions only for a given conformer?
2
A.
Dear Joshua, thank you for your interest and all your questions! Regarding this question: Making prediction for a given conformer will be sufficient.
NS
Nicole Klein Stocke, Community Manager, BASF
January 22, 2025
Is this response helpful?
0
0
Q.
What does the data look like? Is prediction the sole task of interest (i.e. the outcome should ultimately be a classifier/regressor given data labels) or are generative/unsupervised approaches also of interest to study?
2
A.
Data structure needs to be discussed. Initially we would be interested in supervised approaches but are open to suggestions. There is some sort of flexibility in providing those data as they come from numerical simulations.
NS
Nicole Klein Stocke, Community Manager, BASF
January 22, 2025
Is this response helpful?
0
0
Q.
Are solutions with increased speed or higher accuracy of more interest?
2
A.
Solutions with high accuracy are of more interest. However, speed should be kept at least as with classical solutions. Improving both would be the ultimate goal.
NS
Nicole Klein Stocke, Community Manager, BASF
January 22, 2025
Is this response helpful?
0
0
Q.
Which scope of this project would you prefer - focusing on specific spectrum, or developing a general method?
1
A.
Hi Jia-cheng, overall we are interested in general method development as it more broadly applicable however in the course of the project focusing on a small number of spectra is also possible.
NS
Nicole Klein Stocke, Community Manager, BASF
January 7, 2025
Is this response helpful?
1
0
Q.
May I kindly inquire about the typical budget range or ceiling for this project. This will help ensure that the proposed budget aligns with BASF’s expectations and guidelines.
1
Q.
Would you be open to a student capstone course partnering for this project?
1
A.
Hi Kenzo, Thank you for your interest in the QML method call for proposal! Unfortunately, a student capstone project is outside of the partnership scope we are looking for. Thanks for the inquiry!
SS
Sophia Steffens, Innovation & Scouting, BASF
December 18, 2024
Is this response helpful?
0
0
Sign up to access the full partnering request.
View the details of this request and connect directly with corporate R&D teams at Fortune 500 companies.
Share partnering request
Complete
Refine recommendations
Is this request relevant to you?
Eligible partners
Academic researchersStartups
Seeking partners focused on
Analytical ChemistryArtificial IntelligenceBiochemistryComplex SystemsComputational BiologyComputational ChemistryElectrochemistryEnvironmental ChemistryGeneral ChemistryIndustrial Chemistry
20 more

Halo home
Partner smarter. Move faster.
Get new partnering requests
delivered to your inbox.