User login

Navigation

You are here

Comments

Subscribe to Comments feed
Comments
Updated: 18 hours 41 min ago

Open positions at mechanics

Sat, 2022-05-21 06:23

In reply to Postdoc position at Mechanics of composites for energy and mobility group at KAUST

Open positions at mechanics of joints.

Thank you Steve for your very

Thu, 2022-05-19 21:34

In reply to Thank you for the leadership

Thank you Steve for your very comprehensive and thoughtful post! Now we have crossed the threshold where there is more information in the comment section than in the original blog entry :) In response to some of your points: 

1. Thank you for sharing the information on future MMLDT conferences — I attended MMLDT-CSET 2021 virtually last fall, and it was an excellent opportunity to learn more about the field! I am also thrilled to see that the notes from the short course are free to download — that is a very valuable resource. Sharing the Livermore DDPS seminar recordings also reminded me that recordings from the 2020 “Machine Learning in Science and Engineering Mechanical Track” that you and Krishna organized are also available on YouTube: https://www.youtube.com/channel/UCCiwSYhLPtUU3schrt4xviA 

2. Your point about hype vs. pessimism is well stated, and I think the article that you shared “A non-cooperative meta-modeling game for automated third-party calibrating, validating and falsifying constitutive laws with parallelized adversarial attacks” really highlights the importance of challenging our modeling frameworks — both for ML and non-ML based models. I highly recommend that everyone check it out! 

3. I agree — it would be great to see future benchmark datasets focused on multiple types of non-linear mechanical behavior and on challenging microstructures. I think that more access to these types of quite complex data would help advance the development of “mechanics specific” ML methods if/when ML methods that have worked well on simpler problems in mechanics (e.g., Mechanical MNSIT) fail. 

4. Thank you for raising the point about storing trained ML models as well! In addition to being useful for future validation, it is also possible that trained ML models could be useful for transfer learning, although in many cases with mechanical data this may not be straightforward. 

5. Finally, thanks for sharing the link to your lab’s software + data! Just now, I have added the collection of discrete element traction-separation data https://data.mendeley.com/datasets/n5v7hyny8n/1 (manuscript: https://doi.org/10.1016/j.cma.2018.11.026) to the informal list! 

Thank you for the leadership

Thu, 2022-05-19 15:01

In reply to ​​Journal Club for May 2022: Machine Learning in Mechanics: curating datasets and defining challenge problems

Hi Emma, thank you for sharing your vision on this important topic, and for taking the lead to provide the benchmark data with your own time and effort.

Your comment about sharing data and using the same data for benchmarking is spot on. It is almost impossible to create fair and meaningful comparisons of different ML models without a benchmark database. This is complicated by the fact that when we write a paper, we tend to focus on the advantages and promises on the proposed methods, and less so on making the work reproducible and robust, which do not always sound exciting, but is actually very important.

 I think establishing a set of benchmark problems with open-source data for validation and testing can be one step forward to resolve this problem. I also think that open-source the models or at least reporting all the detailed setup required to reproduce the exact results reported in the publications is very important to ensure reproducibility, interpretability, transparency and ultimately the trustworthiness of the proposed method. Without these active measures, it is often difficult to tell whether a model is really doing exceptional well or the product of (intentional/unointentional) cherry-picking . 

I have also attempted to provide my thought on the questions you listed in case it is useful. 

*What resources or upcoming meetings are a good opportunity for others to learn more about the topic? 

The IACM has now introduced a new conference for mechanistic machine learning and digital twins. The first one is at San Diego last year -- https://mmldt.eng.ucsd.edu/home. There will be a second one next year. 

For education resource, thanks to the support of NSF, Professor JS Chen and I have offered a course on the very basic machine learning in mechanics. The videos, lectures, slides, Jupyter notebooks are all free to download. 

https://mmldtshortcourse.weebly.com/lecture-notes.html

There are other colleagues from computer science and in mechanics community that posted great materials. For instance, the Livermore DDPS seminar:

https://data-science.llnl.gov/latest/news/virtual-seminar-series-explores-data-driven-physical-simulations

*For those skeptical about the utility of ML for problems in mechanics in particular (e.g., https://arxiv.org/abs/2112.12054), what would impress you? Can you design a dataset, problem statement, or benchmark challenge problem where ML based predictions would be impactful? 

I believe it is easy to overgeneralize both ways.  There are definitely hypes as well as pessimism extrapolated from small samples of evidence or personal experience. 

There have already been success stories, for instance, in protein folding.  It seems like the difficulty is not to demonstrate some success stories here and there, but establishing universally accepted metrics where different models/approaches/paradigms can be compared and building trust among the modelers/users/stakeholders. 

In the field of constitutive models, we have made a small attempt to build trust by using AI to expose the potential weakness of a given model using reinforcement learning (see below). The idea is to introduce an adversarial agent to explore the loading path and use reinforcement learning to determine the types of loading in which the model tends to perform poorly. Then, this information can be used for re-training such that the weakness can be (potentially) addressed. 

https://www.sciencedirect.com/science/article/pii/S004578252030699X?dgcid=rss_sd_all

I think this can potentially help improving the transparency of the model and avoid cherry-picking via third-party validation.  However, I think having the community to use the same set of benchmark data  (like the Sandia challenge) is probably a better way to move forward. 

*For everyone, what types of benchmark datasets would you like to see in the future? What should new benchmark datasets and associated challenge problems contain? 

I think the dataset you provided is great. I would like to see high-quality data that go beyond elasticity, for example those that involves fracture, damage, twinning, plasticity. Data that involves inverse design (see Kumar, Tan, Zheng and Kochmann 2020) https://www.nature.com/articles/s41524-020-0341-6 for instead and those of interesting microstructures are also great. 

 

*Curating datasets is time, labor, and resource intensive (e.g., see FAIR guidelines https://www.go-fair.org/fair-principles/, https://sites.bu.edu/lejeunelab/files/2022/04/Lejeune_Data_Management_Plan.pdf) — should limited resources (i.e., time, money, storage space) be allocated to these endeavors?

Yes. I think it is necessary. 

*What is the most useful way for mechanical data to be formatted? What necessary metadata should accompany each dataset? 

For practical reasons, data stored in table format is easy to use and share. 

*When does it make sense to curate and preserve data, and when is it unnecessary (e.g., a single FEA simulation can yield GB of data)? 

Whether to preserve the data depends on the opportunity cost and how important of it for the workflow. However, I think in most cases, it is also necessary for the trained model to be preserved as well such that it can be validated in the future if needed. 

*Do you see a role for benchmark datasets in mechanics education? For example, would benchmark datasets be a good resource for a first-year graduate student interested in research at the mechanics/ML interface? What should benchmark datasets for education contain? 

Absolutely. The difficulty is that generating data by itself is very mechanical and the first-year student could be overwhelmed by coursework as well as learning how to do research. 

 

*Do you have a publicly available dataset that we can add to this informal list of curated mechanical data (https://elejeune11.github.io/)? If so, I would love to include it! 

We posted some of our data and codes in our research group webpage and also in Mendeley. 

https://www.poromechanics.org/software--data.html 

 

A proposal to reform the national law for the ASN procedure

Sun, 2022-05-15 23:09

In reply to How to apply to Italian professorship positions -- a little guide

I have written an article with a suggestion on how to improve the ASN procedure in Italy, which has been published to the online journal of the association of ISI high cited italian scientists, which includes Nobel prize winners.  This includes setting higher minimum standard and taking serious control over the abuse of possible "outside sector" outcomes of procedures when foreigners or italians abroad apply. 

https://www.scienzainrete.it/articolo/riforma-della-abilitazione-scientifica-nazionale/michele-ciavarella-vito-dandrea/2022-05-14 

Cartesian vs Natural

Sun, 2022-05-15 17:40

In reply to How to evaluate natural coordinates from global coordinates?

You must determine the Jacobian from x(r,s,t) , y(r,s,t) and z(r,s,t) where x(r,s,t)=ΣNi(r,s,t)*xi

One has to find the Ni 8-nodes shape functions for example if node 1 is located at r=-1 , s=-1 and t=-1 then all the other nodes are at r=1 (ie 1-r=0) , s=1 (ie 1-s=0) or t=1 (ie 1-t=0) one can find

N1(r,s,t)=(1-r)(1-s)(1-t)/8 This polynome must scan all the other nodes. If one replaces the node 1 coordinates, one must find the value N1=1 at this node the division by 8 allows this. N1=0 at the other nodes.

x,r=ΣNi,r*xi in the Jacobian. The same formulation can be applied to the other nodes , to x,s , to x,t ,to y,r , to y,s , to y,t , to z,r , to z,s and  to z,t . The element matrices use the shape functions derivatives in their integrals which can be computed with the gaussian quadrature formulas.

my resignation has been revoked !!!

Sun, 2022-05-15 00:56

In reply to How to apply to Italian professorship positions -- a little guide

Due to high pressure to remain in the National commission, and resulting small scandal if I had to resign, I have been asked to withdraw the resignation, and I have accepted to remain in the commission.

 

I highly recommend any potential candidate, to approach me for any clarification by email mciava@poliba.it 

 

Regards

MC 

A facebook group of 14.500 members will help you in applying!

Thu, 2022-05-12 21:27

In reply to How to apply to Italian professorship positions -- a little guide

Please subscribe to the facebook group https://www.facebook.com/groups/166806536849115    to learn about ASN, make questions, and get to know each other with 14500 members.  A group I created in 2012, so now is 10 years old.

regards, MC

Thanks you Francisco for your

Tue, 2022-05-10 20:48

In reply to Thanks Emma for the very

Thanks you Francisco for your thoughtful points! In response: 

1. Yes! I look forward to seeing future creative approaches and perhaps more generalizable insights that are enabled by broader access to mechanics-based datasets. Hopefully data sharing can increase synergy between researchers with different expertise. 

2.  Thanks in particular for raising this point! And I completely agree — there are so many different and innovative ways to leverage both a fundamental understanding of mechanics + creative ideas for modifying open source ML software that are less costly to implement than the initial data generation step. 

3. Thanks for sharing this benchmark dataset! For others who may be interested, the dataset is hosted through the “Cardiac Atlas Project” http://www.cardiacatlas.org/ which has a specific “Motion Tracking Challenge” http://www.cardiacatlas.org/challenges/motion-tracking-challenge/

Finally, I very much look forward to seeing more work from your group on research at the intersection of imaging + machine learning + mechanics!

Thank you Jessica for your

Tue, 2022-05-10 20:29

In reply to Thanks Emma for such an insightful discussion

Thank you Jessica for your kind comments. I hope that your students also find the post helpful! There is so much exciting research going on with applying ML to mechanics, and so many opportunities for people to contribute new ideas!

correction to step b

Tue, 2022-05-10 19:57

In reply to global to natural

deltaRn=-inverse(J)*Rs

global to natural

Tue, 2022-05-10 19:55

In reply to How to evaluate natural coordinates from global coordinates?

The basic Newton-Raphson scheme can be used to get the corresponding natural coordinates.  There is no direct equation (that I am aware of).  You do need to solve a small nonlinear equation (Newton-Raphson iterations).

Here is the basic outline:

0. Set the global point you are interested, in create column vector Xg=[xg; yg ;zg]

1. Set column vector of initial guesses for Rn=[rn;sn;tn]

2. At the current values for rn,sn,tn evaluate column vector X=[x(rn,sn,tn); y(rn,sn,tn); z(rn,sn,tn)]

3. Calculate the norm of residual(error), |Rs|=|Xg-X|, compare to allowed tolerance, perhaps you want |Rs|<10^-6

4. While tolerance not met iterate until tolerance met

a. At current value (or guess) for rn,sn,tn find the Jacobian, J(rn,sn,tn)=[dx/dr dx/ds dx/dt; dy/dr dy/ds dy/dt; dz/dr dz/ds dz/dt]

b. Determine correction for Rn,   deltaRn= -J-1Rs

c. Calculate column vector Rn+1=Rn+deltaRn      =        [rn+1;sn+1;tn+1]

d. Calculate new residual, use new X at values [rn+1;sn+1;tn+1], then Rs=Xg-X

e. Repeat until desired norm of residual |Rs| is below tolerance

End iterations

Rn+1 will contain your r,s,t values, that correspond to global values Xg

 

I think I have the above algorithm worked out properly.  Hopefully, it at least gets you headed in the right direction.

 

You may also like to look at the following links

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4991307/

https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.971.6670&rep=r...

 

 

Thanks Emma for the very

Tue, 2022-05-10 18:14

In reply to ​​Journal Club for May 2022: Machine Learning in Mechanics: curating datasets and defining challenge problems

Thanks Emma for the very insightful post!

I agree with you that just releasing these datasets fosters discovery. Just by looking at the description of the datasets we can come up with novel machine learning techniques to address that particular problem.

I would also like to point out that the availability of these datasets levels the playing field for researchers from universities with fewer resources, which may not have access to supercomputers to run thousands of simulations or to precise experimental setups. Initiatives like this may increase the pool of researchers interested in the intersection of mechanics and machine learning, which can only be beneficial for the field.

Finally, I would like to mention a benchmark dataset in the field of cardiac strain estimation from different imaging modalities: https://doi.org/10.1016/j.media.2013.03.008. This benchmark has been used by many other researchers and has become the gold standard dataset to compare algorithms in cardiac image registration. Even though many applications are directly related to imaging, there are incredible opportunities at the intersection of imaging+machine learning+mechanics, some of which we are working on!

Thanks Emma for such an insightful discussion

Tue, 2022-05-10 17:02

In reply to ​​Journal Club for May 2022: Machine Learning in Mechanics: curating datasets and defining challenge problems

Thanks Emma for posting such an insightful discussion on this new, exciting research topic! Many researchers started to use ML in their research, but what are the new challenging problems and opportunities?  Your post provides a very thorough overview of ML in mechanics, particularly focusing on curating datasets, and also answers to these questions. This is a great resource with many details for young people who want to jump into this emerging research area. I will share your post with students in my lab and my teaching classes at Carnegie Mellon.

Alain Goriely Elected as FRS & Springer Lectures

Tue, 2022-05-10 09:38

In reply to Alain Goriely's Seminars on Growth at UC Berkeley

Dear Colleagues,

 Please join me in congratulating Alain Goriely on his election as a Fellow of the Royal Society: https://royalsociety.org/news/2022/05/new-fellows-2022/  Having Alain at UC Berkeley as a Springer Professor these past four weeks has been a great pleasure.  His last two lectures will be held today and Thursday (zoom access available - see links above).  Best wishes,Oliver

 

 

I have resigned from the ASN commission in protest

Mon, 2022-05-09 12:12

In reply to How to apply to Italian professorship positions -- a little guide

Communication to friends and colleagues sector Ing Ind 14

 

With great regret, I must report that I have to give up my recent evaluation activities. I put the utmost commitment and my ability, including diplomatic, into carrying out this task, which I considered and still consider to be of the utmost seriousness and importance. However, from the very beginning, provincial, kamikaze, and irregular logics prevailed in others (if they can be called logical, since they are not rational), so weak candidates who cannot even write in English with regular grammar should be favored. , rather than candidates also coming from abroad and prestigious Universities, which have published in the best magazines in the sector. I tried to understand this logic until the end, and I was not alone, but indeed we were 2 out of 3 of the SSD thinking like me, however this third party continued to bully, rant, make agreements with minor SSDs under the bench, insult, threaten and speak scurrilously and vulgarly. Now I can't take it anymore and I have to say enough. I cannot tarnish my name with these processes, on the contrary I have reported them to the Ministry and we hope they will be punished. Sorry everyone for any temporary damage this will bring, but I am sure that anyone will do better than the team that was created. The late Professor Lazzarin, my great friend, one day before his death asked me to continue to do well for our SSD, which is not very modern in fact it is in crisis, and therefore must open to the outside. And to honor his memory, I would never have been able to accept certain stomach-churning results, bordering on the nepotistic and parish interests, indeed beyond this limit.

I am available for clarification in person.

HOWEVER; GOOD CANDIDATES SHOULD APPLY WAITING FOR THE NEW COMMISION!

 

Michele Ciavarella

Congratulations on this

Fri, 2022-05-06 10:27

In reply to Water as a “glue”: Elasticity-enhanced wet attachment of biomimetic microcup structures

Congratulations on this fantastic work!

I am just curious what will happen if the substrate is nonflat? For example, cylinder or sphere surface?

Another question is, what if the substrate is also deformable?

III quarter deadline approaching to apply !!!

Fri, 2022-05-06 09:47

In reply to How to apply to Italian professorship positions -- a little guide

III quarter deadline approaching to apply to associate or full professorship habilitation in italy

 

III quarter: starting from February 2, 2022 and no later than 00.00 (Italian time) of June 3, 2022;

 

You can contact me for any details

 

MC

 

Yes! Thanks again for sharing

Fri, 2022-05-06 08:45

In reply to I am glad to know that you

Yes! Thanks again for sharing it — it is super relevant to this topic! 

With regard to following a template, I have four comments:

1. Because we are based at Boston University, we have been using the OpenBU Institutional Repository (https://open.bu.edu/). For each submission, we follow the OpenBU template that includes components such as a thumbnail image, an abstract, data rights, a hierarchy of dataset “collections,” and links to the relevant code (see attached figure).

 

2. Broadly speaking, we have been guided by trying to adhere to FAIR Principles (https://www.go-fair.org/fair-principles/). 

3. Thus far, the scope of our work is relatively small (i.e., we share medium sized computationally generated datasets where researchers can quickly download input files and output files to use for training ML models). Therefore, formatting these particular datasets is much less of a challenge than it could be for mechanics data broadly defined. 

4. For one of our recent datasets (Mechanical MNIST Crack Path) we actually ended up publishing two versions of the dataset: a “lite” version (https://open.bu.edu/handle/2144/42757) that matches the format of other datasets in the Mechanical MNIST collection, and an “extended” version (https://datadryad.org/stash/dataset/doi:10.5061/dryad.rv15dv486) that offers much more flexibility, at the expense of a slightly higher barrier in getting started. 

 

Do you know of any additional resources that are useful in this direction? In addition, I am also curious if you (or others!) have thoughts on the accessibility vs. flexibility tradeoff in data curation mentioned above. 

I am glad to know that you

Fri, 2022-05-06 01:08

In reply to Thank you Ajay for this very

I am glad to know that you found the DesignSafe database useful. Yes, it has a lot of data from civil, structural and coastal engineers. This includes data related to experiments and computations. These are some nice ones that you have identified and I am happy to see that there is also a contribution from someone from Auckland!

Do you have a template that you ask the students to follow when creating these datasets? It would be particularly important to have a standard way of doing these to ensure uniformity across them.

 

Thanks so much Manuel for

Thu, 2022-05-05 16:49

In reply to Leading by inspiration

Thanks so much Manuel for your kind words! And thank you for sharing the fantastic work that you have been doing in your lab on making open access mechanics datasets. Three brief follow ups: 

1. I want to re-emphasize your point on the investment of money and time that goes in to collecting these experimental datasets — even “large” experimental datasets like the ones that you have shared are relatively small compared to what is available for “big data” in other fields. Overall, I think ML methods that can leverage these small high quality datasets (perhaps in conjunction with standard simulation methods) are quite relevant to the mechanics field.

2. I also think it’s really great that in addition to providing these data, you and your team have invested significant additional effort in making these datasets accessible to others through documentation (e.g., https://dataverse.tdl.org/file.xhtml?fileId=105543&version=1.0). 

3. Finally, I want to point out that you made this data public through the “Texas Data Repository” (https://dataverse.tdl.org/). It seems that this is a great resource for others who are affiliated with universities that are  Texas Digital Library (TDL) member institutions (https://dataverse.tdl.org/). 

Pages

More comments

Syndicate

Subscribe to Syndicate