Changes between Version 2 and Version 3 of CodeCampIX/BreakOut1

Mar 31, 2019 11:01:49 AM (14 months ago)


  • CodeCampIX/BreakOut1

    v2 v3  
    2424    * Project description 
    2525    * Use cases covering broad base of different **types** of problems 
     27== Minutes == 
     29=== Reparameterisation of models === 
     30 * Use Cases: 
     31  * Amounts of water in outer layer vs SLD of layer 
     32  * Number of particles / packing e.g. Volumes of heads and tails tied 
     33  * Examples of complicated things from Sylvain and Richard e.g. micelles with charge and molecular constraints 
     34 * How to implement? 
     35  * Make lots of models and use categories to manage. Each parameterization of a model is now its own model 
     36  * Provide a "workflow" that uses the contraints system to create on the fly 
     37  * Create new infrastructure to add calculated parameters on the fly 
     38 * How do we do this? 
     39  * Can we make a good interface for the user to do parameter mapping  
     40  * Wojciech - the optimiser might have problems with that. 
     41  * Non-fittable parameters 
     42   * Involved with a constraint 
     43   * Calculated outputs 
     44   * See SLD profiles, see 3D shape 
     45  * Narayan: Make sure that it is clear that constrained parameters are visible and clear that they are constrained. 
     46  * PDB: Actually should be easy to provide an on the fly reparameterization: Use a GUI tool like the Add Model Editor to provide parameters, any constraints and their relationship to the existing parameters.  Then "create the code" by editing the model file: adding the conversion math at the beginning of the function. Essentially the infrastructure and a first order proof of concept already. 
     47  * Start with a simple test, then maybe talk to Paul Kienzle 
     48  * Think about GUI 
     50=== Include more complex structure factors === 
     51 * Discussion of the implementation of F(Q) 
     52  * We don’t have complex amplitudes 
     53   * Use case : Janus particles - but can cheat it with end capped cylinder 
     54 * Beta 
     55 * Locally monodisperse 
     56 * More ways of dealing with structure factors 
     57  *pyPRISM - will be implemented in SasView by Tyler Martin 
     58 * Discussion of SasFit / SasView 
     59 * Models 
     60 * OZ Solver 
     61 * More ways of dealing with polydispersity 
     62 * Orientation dependent Structure factors that depend on form factor (amplitude) 
     63Write analytical ones 
     64Yun - only good for hard sphere and spherical particles  
     65Yun - Continuous distributions in OZ framework is hard 
     66Yun - should go for numerical solutions for more complex interaction potentials 
     67Yun - for e.g. mixture of latex and gold, need cross terms, so need complex amplitudes 
     68 * Sylvain - Finish decoupling approximation first? For spherical. 
     69ALL - AGREED last point 
     71=== Include material from Scatter === 
     72Andrew has had student working on it. Difficult. Needs hypergeometric functions. 
     73MLZ have been working on it too - a few simple models. 
     74They have some proof of concept in Born Again 
     75Need to work out API to work with SasView 
     76Plan for this ... 
     77Aim for next Code Camp to have BornAgain team integrate their code. 
     78Check SasFit - Joachim may have implemented already. 
     80=== 1D oriented data ... === 
     81How to fit a slice from a 2D dataset 
     82Do we do it? How? 
     83Make special 1D models? 
     84Or cut 2D models? 
     85Speed … ? Need to look at the speed ...= 
     86User interface … ? 
     87How to show the slicer options to the user? 
     88Narayan / Sylvain to find some example data. 
     89Maybe start with simulation and treat as real data to test process 
     90AJJ to make some simulated data and share with everyone. 
     92=== Project to do benchmarking of optimisers === 
     93* Wojciech : Idea is that there is a summer student at ISIS working with Anders 
     94Mantid have been testing their optimisers seeing how they perform on different classes of problems.  Want to do something similar for SasView 
     95 * Test SasView optimisers against their defined problems 
     96 * Identify classes of problems that would be representative for the SANS community 
     97* AJJ - need to know the answer? 
     98* AJJ - how is SANS mathematically different from other test cases? 
     99* RKH - What about things where we have models that have multiple possible simulations? 
     101=== Computation speed and integration options and approaches === 
     102* Ingo : Suggestion - improve the integrations inside models. 
     103 * Get rid of nested integrals 
     104 * Use lebchev(?) Integrals? - some unrolled already- but other would require balancing competing requirements? 
     105 * How to include with GPU code ? 
     106 * Testing already … some models have numerical integration problems. Work underway. 
     107 * Richard - will always have convergence issues for large particles. Or at least with particles that have one dimension that is massively larger than the other. 
     108* Hypergeometric functions 
     109 * Can we implement them for more models? 
     110 * AJJ to find code from Johan.