Welcome to the ML for ssNMR Workshop in Manchester! We are excited to have you join us. This page contains important information for your participation.

This page has practical information for attendees, including the programme, travel information, and accommodation details. If you have any questions, please feel free to reach out to the organisers.

Workshop Details

* Note that meals and accommodation at the conference hotel for the night of 13th May are included for registered attendees. This was enabled thanks to generous funding of the CCP-NC via the UKRI DRI grant.

Programme - TBC

Your discussion suggestions

Below are some common themes amongst the suggestions you made during registration. We will be using these to guide the discussions during the workshop.

1. Transferability of ML Techniques

• Exploring the applicability of interatomic potential ML methods (e.g. GAP/SOAP/MACE) to ssNMR property prediction.
• Can force field models be adapted for accurate prediction of NMR observables?

2. Integration with DFT and Advanced Methods

• Strategies for interfacing DFT codes (e.g. CASTEP) with ML-based models.
• Challenges in linking DFT-computed parameters to experimental ssNMR data.
• Incorporating non-local and beyond-DFT corrections into ML workflows.

3. Finite-Temperature and Dynamic Effects

• Methods for capturing finite-temperature effects in NMR spectra using ML.
• Role of ML force fields in enabling long-timescale molecular dynamics.
• Translating computed molecular motions into measurable ssNMR parameters.

4. Machine Learning for MOFs and Guest Molecules

• ML-guided modelling of MOF structures and dynamic guest-host interactions.
• Predictive modelling of ssNMR parameters for porous materials.

5. Data Quality and Reference Accuracy

• Impact of noise and uncertainty in reference DFT/chemical shielding datasets.
• Techniques for mitigating sensitivity to noisy training data in ML models.

6. Benchmarking and Standardisation

• Establishing standard datasets and evaluation protocols for ML-ssNMR models.
• Ensuring reproducibility and robustness of ML workflows.

7. Relevance to Experimentalists

• Identifying experimental requirements that ML-ssNMR must address.
• Enhancing accessibility and interpretability of ML tools for ssNMR practitioners.

8. Diversity and Uncertainty Management in ML Models

• Exploring diverse ML model architectures suited to ssNMR prediction.
• Techniques for uncertainty quantification and confidence calibration in ML predictions.

Posters

We will have posters set up throughout the two days with ample time for discussion between sessions. A PDF of the submitted poster abstracts can be downloaded here: ML for NMR Workshop Poster Abstracts.

Travel Information

Location: Hyatt Regency Manchester, 55 Booth Street West, Manchester, M15 6PQ

The hotel has a nice breakdown of the travel (and parking) options available on their website: Hyatt Regency Manchester - Directions.

Accommodation

Accommodation for the night of Tuesday, 13th May 2025 at the conference hotel (Hyatt Regency Manchester) is included for registered attendees (unless you indicated you didn't require that, of course). If you require accommodation for additional nights, please arrange this directly with the hotel or other local providers.

Meals

Lunch and dinner on Tuesday, 13th May 2025 are included for registered attendees. Breakfast on Wednesday, 14th May 2025 is also included. Any dietary requirements should have been indicated during registration; these have been passed on to the venue. If you have any last-minute changes, please do let us know as soon as possible.

Tea and coffee etc. will be provided during breaks.

Hybrid/Recording

Unfortunately, due to the emphasis on discussion during this workshop, we cannot offer a hybrid/recording option. However, we will be publishing a report summarising some of the overarching themes on the CCP-NC website after the event.

Contact

If you have any questions regarding the workshop, please contact one of the organisers:

• Dr Albert Bartók-Partay
• Dr J. Kane Shenton