Learning from the Harvard Clean Energy Project: The Use of Neural Networks to Accelerate Materials Discovery

By Edward O. Pyzer-Knapp, Kewei Li, Alán Aspuru-Guzik. 2015

205 citations recorded by the source corpus.

Subfields: Catalysis, Computational Theory and Mathematics, Materials Chemistry, Computational Chemistry & Drug Discovery, AI for Physical Sciences

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Catalysis

1. José Miguel Hernández-Lobato
2. Bo Dai
3. Le Song
4. Hanjun Dai
5. Alán Aspuru-Guzik

In Computational Theory and Mathematics

1. José Miguel Hernández-Lobato
2. Thang D. Bui
3. Florian Hase
4. Alán Aspuru-Guzik
5. Pascal Friederich
6. Le Song
7. Hanjun Dai
8. Jonny Proppe

In Materials Chemistry

1. José Miguel Hernández-Lobato
2. Kevin Maik Jablonka
3. Seyed Mohamad Moosavi
4. Berend Smit
5. Florian Hase
6. Pascal Friederich
7. Thang D. Bui
8. Alán Aspuru-Guzik

In Computational Chemistry & Drug Discovery

1. José Miguel Hernández-Lobato
2. Florian Hase
3. Kevin Maik Jablonka
4. Berend Smit
5. Le Song
6. Bo Dai
7. Hanjun Dai
8. Seyed Mohamad Moosavi

In AI for Physical Sciences

1. José Miguel Hernández-Lobato
2. Bo Dai
3. Le Song
4. Hanjun Dai
5. Alán Aspuru-Guzik
6. Jonny Proppe
7. Pascal Friederich
8. Florian Hase

Citations in the corpus

• Discriminative Embeddings of Latent Variable Models for Structured Data (2016, positive): “irection for the high throughput screening and discovering new materials. In this experiment, we randomly select 90% of the data for training, and the rest 10% for testing. This setting is similar to Pyzer-Knapp et al. (2015), except that we use the entire 2.3m dataset here. Since the data is distri”
• Deep Gaussian Processes for Regression using Approximate Expectation Propagation (2016, neutral): “16 based on a suggestion in (Pyzer-Knapp et al., 2015).”
• Black-box $\alpha$-divergence Minimization (2015, positive): “The noise variance is fixed to 0.16, which is the optimal value according to the results reported in Pyzer-Knapp et al. (2015).”
• Machine learning exciton dynamics (2015, unknown): “This demonstrates the power of machine learning in chemistry, as has also been found in recent work where neural networks were employed to extract other chemical properties [40–44] .”
• Big-Data Science in Porous Materials: Materials Genomics and Machine Learning (2020, methodology): “epth and another viewpoint in section 6) which acts as a high-pass filter, i.e., it filters out the noise and makes the inversion more stable and the solution smoother. One then solves a = (K +lI) 1 y. (16) The most widely known algorithms which use this kernel trick are support vector machines (SVMs)”
• Variational Implicit Processes (2018, background): “Previous studies expert-designed [11, 37, 71] or learned finger-print features [25] that transforms the raw string data into fixed-size feature for prediction.”
• Machine-learned potentials for next-generation matter simulations (2021, result): “In contrast to many other machine learning tasks in materials and molecular sciences 49,83–85 , off-equilibrium conformations are therefore of crucial importance when training ML potentials.”
• Neural Networks for the Prediction of Organic Chemistry Reactions (2016, unknown): “While methods that are guided by quantum calculations have the potential to explore a wider range of reactions than the heuristic-based methods, these algorithms would require new calculations for each additional reaction family and will be prohibitively costly over a large set of new reactions.”