We are research group @SecondmindAI. We tweet about probabilistic modelling and active learning algorithms, and about tackling real-world problems with them.
Last week @mutny_ml (ETH Zurich) gave talk at our research seminar on "Optimal Experiment Design in Markov Chains". We have a learned a lot!
Details and recording available at: https://t.co/bPmtvqrp0E
We are currently hiring a Machine learning researcher to join the Labs: https://t.co/ukYP0BJoSF
If you are into probabilistic modelling or active learning, and care about using these methods to solve real-world problems, Secondmind is the place to be :)
We are currently hiring a Machine learning researcher to join the Labs: https://t.co/ukYP0BJoSF
If you are into probabilistic modelling or active learning, and care about using these methods to solve real-world problems, Secondmind is the place to be :)
In the meantime, our open-source Bayesian optimization toolbox built-on @TensorFlow - Trieste - has matured and now stands at 3.0.0: https://t.co/EpEOYtJFRX
Do give it a check if you need a flexible yet robust active learning library :)
And @ToyotaMotorCorp work on making diffusion models more realistic looks full of potential. Check their paper on drag-guided diffusion model for car design: https://t.co/KvyNFX47pR
On a slightly different topic, we have recently visited a conference on Generative AI for automotive industry: https://t.co/uCjGK2THth. We saw some really cool work, exciting products and features are in the works!
Our schedule for new seminar season has slowly started filling up: https://t.co/5KMYTwTdth. We are looking forward to this week's talk by @mutny_ml from ETH Zurich, on "Optimal Experiment Design in Markov Chains".
Our last seminar was back in 2023, when @Domenic_DF gave a great talk on Bayesian modelling and decision making in civil engineering, do check it out here: https://t.co/6rRLHEG6hY.
Here is a translation to ML terms: We are solving a multi-output level set estimation (or feasible set identification) problem, and scaling on all fronts - from data size and input dimensionality, to generating hundreds of recommendations in a batch.
We have been busy working on some basic research, as well as developing technology behind our products. In particular, we have been busy with our new System Design product: https://t.co/M0ixG4Jht5
For companies relying on costly high-dimensional simulations to manage design of complex products (e.g. cars, trucks or other vehicles), this is an active learning based solution that will make design exploration more efficient and (hopefully) eliminate costly rework.
We are delighted to have been recognised by @AutoTechAwards as the winner of its AutoTech AI Innovation of the Year 2023 Award.
👉 Read more here: https://t.co/39UZQN3orO
Our #ICML2023 paper introduces Neural Diffusion Processes - generalising diffusion processes to function spaces.
Great to work with @vdutor@alandanielsaul and @ZoubinGhahrama1
Paper: https://t.co/gXhIQ2xRFF
Code: https://t.co/6zHgZb8VN7
We’re looking for a Machine Learning Researcher to join our award-winning team. If you're looking for a new challenge, you can find out more and apply here 👉 https://t.co/hT0VkQC8xs
Instead of sending cards this Christmas, our team has chosen to make a donation to @TrussellTrust, a network of food banks across the UK that provides emergency support to families at crisis point.
You can find out more about their work here: https://t.co/Ja1hSp6zte
The second edition of the NeurIPS @ Cambridge meetup is tomorrow - and it's a sell out! Our team will be there to meet other researchers in the local ML community, and soak up the panel sessions and presentations. Hope to see you there! https://t.co/GJIJx5rAxO