#machine-learning
2019-11-28
Dario Strbenac (19:00:04): > @Dario Strbenac has joined the channel
Dario Strbenac (19:00:07): > set the channel description: Bioconductor-centric Bioinformatics-flavoured Machine Learning
Vince Carey (19:00:34): > @Vince Carey has joined the channel
Alan O’C (19:00:35): > @Alan O’C has joined the channel
2019-11-29
Vince Carey (06:17:37): > Thanks@Dario Strbenac. A number of references to tensorflow have come up in Bioc development, and@Aaron Lunhas created the basilisk package to help systematize Bioc-python interfaces. Perhaps a plan for simplifying use of tensorflow with bioc could be sketched in this channel (among other plans…)
Aaron Lun (06:17:43): > @Aaron Lun has joined the channel
Aaron Lun (13:30:40): > basilisk now pulls down anaconda on install, with the specified version of tensorflow. Seems to work well withhttps://github.com/Irrationone/cellassign/issues/36. Ready to get into BioC once I get some assistance on issues inhttps://github.com/LTLA/basilisk/issues/1.
Dror Berel (17:42:23): > @Dror Berel has joined the channel
2019-11-30
Dario Strbenac (00:00:15): > Ah, I haven’t used TensorFlow. Has anyone seen any scenarios where the new and trendy classifiers perform significantly better than older classifiers popular in R for bioinformatics classification problems (largep, smallndata set)? I often see tables in machine learning journal articles where they show their method has 1% or 2% better performance than an existing method and they report their performance metric to three or more decimal places. If an oncologist or cellular biologist has a data set of 50 cancer patients (a typical size if you’re not the TCGA consortium but rely on typical research grants), they don’t get excited if a classifier is developed that has 78.456% accuracy instead of 76.123%. I work in a statistics department that frequently collaborates with biologists and the statisticians cringe when someone uses the phrase “machine learning”.
Dario Strbenac (00:05:09): > A co-worker of mine recently developed an interesting bioinformatics classifier that works surprisingly better than its componentshttp://www.oncotarget.com/index.php?journal=oncotarget&page=article&op=view&path%5B%5D=13203&path%5B%5D=41874The basic idea is: > 1. Do repeated cross-validation on gene expression data within the training partition of samples. Calculate a per-sample error rate (i.e. how often each patient is misclassified). > 2. Search the matched clinical data for clinical variables which explain why some patients are tough to classify correctly using omics data (i.e. regression of covariates on the patient error rate). > 3. Create a rule like IF extranodal_spread IS Yes THEN Poor Outcome ELSE predict either Poor or Good Outcome using trained gene expression classifier (such as diagonal LDA) and apply to test set. > Repeat many times to get estimate of performance variability. > > Such a two-stage classifier works better than putting all of the clinical and omics variables into a single classifier and also better than Tibshirani’s pre-validation approach to combining the largepomics data with the smallpclinical data (Figure 2). > > I think this style of classification that is specifically tailored to the problem is what sets ClassifyR apart from a package like caret. ClassifyR has a version of the aforementioned approach that avoids the nested cross-validation and other functions for gene network-based classification.
Dario Strbenac (23:30:19): > I had time today to explore mlr. It is comprehensive and I notice many similarities to ClassifyR. One important aspect of Bioconductor is that the user base is largely comprised of statisticians. Many of my colleagues who have tried ClassifyR didn’t like it. Its design uses a lot of S4 objects, constructors and accessors which they find overbearing. I studied computer science when I was an undergraduate but my colleagues are statisticians. In the design of the new Bioconductor classification framework, how could it be designed to appeal to statisticians?
Dror Berel (23:46:04): > To me, the main benefit of using specialized packages from other domain, is that I don’t need to reinvent the wheel. For example, it first drove me crazy that no place across the entire rich package of mlr, p-values are never mentioned. Not even in a simple use of linear regression. However, I would still use it for it’s rich use of sampling (e.g. cross validation) and other solutions. To me, the fun part is merging such domain specialized packages together, and knitting them together
Dror Berel (23:48:04): > By the way, mlr is about to be deprecated, refactored into mlr3
2019-12-01
Aaron Lun (00:29:22): > @Dario StrbenacI find it helpful to have the same functions be applicable on both base R objects and S4 equivalents. For example, many of my single-cell packages can accept a SingleCellExperiment or a base matrix, so people are never forced to use the former if they just want to get a job done. It is often the case that some extra stuff can be done with the richer object, so I also provide functionality for that, but people can just use the basics if they like.
Dario Strbenac (01:30:07): > I like to do the same. Thematrixvariety transposes the matrix (so the variables are in columns) and casts it toDataFrame. TheMultiAssayExperimentuseswideFormatto convert the input data into aDataFrame. TheDataFramevariety is the only variety which does the computations. Do the statisticians in your workplace use any meta-classification package such as mlr and how do they find the user experience?
Aaron Lun (01:48:26): > hold on, so why are they complaining about constructors, then?
2019-12-02
Dario Strbenac (15:00:08): > Object-oriented programming isn’t covered in a statistical degree, I think.
Aaron Lun (15:02:20): > Well, I would imagine that they wouldn’t even be able to care about those objects if the methods support direct use of base R things.
Dario Strbenac (18:00:10): > I found that they baulked at the complexity / size of the parameter list of the function which runs the cross-validation and the number and diversity of slots of the classification result object. One idea I had was to put a Shiny front-end over it, but I never had the time to implement it. Event-driven programming adds a lot to a software’s development time (particularly if you don’t practise it regularly) and no reward from academic employment.
2019-12-03
Alan O’C (04:52:43): > I think that’s a valid thing to baulk at; common convention is that > 8 or so parameters is too many for end users and designers alike. Also given that R is functional, users probably shouldn’t have to dig into slots too often
2019-12-04
Mikhael Manurung (22:23:12): > @Mikhael Manurung has joined the channel
2019-12-08
Dror Berel (11:05:44): > Just completed my first draft of a blog post about the softwares design differences between mlr3 and tidymodels. Hope you will enjoy reading this less-academic light post, and please feel free to share your thoughts about ithttps://medium.com/@drorberel/meta-machine-learning-aggregator-packages-in-r-round-ii-71ee1ff68642 - Attachment (Medium): Meta Machine Learning aggregator packages in R: Round II > written on January, 2020.
Dario Strbenac (21:00:03): > It’s interesting to know about what is coming soon. Those projects seem to have a lot of people involved and will have lots of complex features.
2019-12-09
Alan O’C (05:46:38): > Not sure I agree that mlr3 will be a more restrictive framework. In my experience tidyverse packages are more than happy to tell you to jog on if you try to do something they didn’t exactly prescribe:slightly_smiling_face:
2019-12-10
Sebastian Gibb (12:07:15): > @Sebastian Gibb has joined the channel
2019-12-11
Elisabetta Mereu (06:09:27): > @Elisabetta Mereu has joined the channel
2019-12-17
Jean Yang (02:39:48): > @Jean Yang has joined the channel
2019-12-31
Leopoldo Valiente (18:20:11): > @Leopoldo Valiente has joined the channel
2020-02-08
Thanh Le Viet (16:39:55): > @Thanh Le Viet has joined the channel
2020-02-13
Noor Pratap Singh (00:21:07): > @Noor Pratap Singh has joined the channel
2020-02-14
Andrew Skelton (05:07:04): > @Andrew Skelton has joined the channel
2020-02-22
Aedin Culhane (07:42:43): > @Aedin Culhane has joined the channel
2020-02-26
Jamie Burke (15:46:25): > @Jamie Burke has joined the channel
2020-03-10
Aaron Lun (02:58:59): > @Vince Careybasilisklives! Ready forBiocsklearnmigration. Note that if you only depend on scikit-learn, you don’t even have to set up your own conda environments; just setNULLinbasiliskStartto use the base environment that gets installed withbasilisk.
2020-03-19
Daniela Cassol (12:31:35): > @Daniela Cassol has joined the channel
2020-03-25
Aaron Lun (23:45:18): > @Vince CareyDo you think you can flip Biocsklearn to basilisk before the release? I’d like to get a few clients to test it out “in the wild”.
2020-03-26
Vince Carey (07:43:42): > i want to@Aaron Lun…let me try this morning.
Vince Carey (08:05:07): > > # The return value MUST be a pure R object, i.e., no reticulate > # Python objects, no pointers to shared memory. > > I take it from this that it defeats the purpose of basilisk entirely if I want the user to be able to manipulate python references after running BiocSklearn code.
Aaron Lun (11:53:49): > You can continue to do so but only via basiliskRun.
Aaron Lun (12:34:40): > So for example, you can create a persistent variable that holds a python reference and just use it in another basiliskRun call to the sameproc. See?basiliskStartandfindPersistentEnvfor more details.
Aaron Lun (12:39:46): > I’m not entirely happy with how I’m passing down the persistent envs, so I may fiddle with this to make it easier to use.
2020-03-31
Aaron Lun (02:09:38): > @Vince Careylooks like we’re halfway there. Windows failing again, what a surprise.
Aaron Lun (02:16:57): > Also, theop <<-is not advisable. If you want to persist something across functions, you would have to hold ontoproc;opis not guaranteed to exist onceprocis stopped.
Aaron Lun (02:27:46): > The nature of the failure provides little insight, I’m afraid. It would seem that (i) somehow Python was already loaded into the R session when the vignette compiles, causing basilisk to create a new process on a separate node; and (ii) the new process fails when it loads the wrong conda installation, heaven knows why.
Aaron Lun (02:29:12): > If you have access to a Windows machine to debug, I would be very interested to know the cause of this.
Vince Carey (08:25:44): > i should have some time today to look at this. i have windows running under virtualbox on a mac.
kipper fletez-brant (11:44:52): > @kipper fletez-brant has joined the channel
2020-04-02
Aaron Lun (03:53:59): > @Vince CareyIf you find anything, let me know. The windows build failures have infectedbasiliskitself, with the standard level of unbelievability.
Vince Carey (06:50:22): > OK – too many calls today but it is high on the list
Vince Carey (21:39:41) (in thread): > @Aaron Lunwe’ll have to discuss this more .. I think if I get it right, the whole BiocSklearn API can be preserved (I won’t have to rewrite vignette) but the persistent env to support reuse of python objects has to be done correctly and durably. These objects give us access to python help via py_help and I think that is useful.
Aaron Lun (21:51:54) (in thread): > How one would theoretically do this is to pass theprocaround as the input/output of the various functions. This is potentially nasty to the user asbasiliskwill clearPYTHONPATHin betweenbasiliskStartandbasiliskEnd. (It has to do so, otherwise Python wanders off and fetches the wrong versions of packages.) As a result, if you have a function that callsbasiliskStartbut does not callbasiliskEnd, the R session is left in a state of not playing nice with user’s Python code. Which isn’t a big deal to me, but might be frustrating to some people. > > In package development contexts, any other developer calling your functions needs to know that they have to also callbasiliskEnd(or a wrapper around such a function) to terminate the process. Otherwise you get a bunch of errors inR CMD checkwhen it detects open processes in Examples. > > There is some room for writing functions that return a livebasiliskprocess, but those functions should be limited to “advanced use”, given the caveats mentioned above. For general purpose use, it would be best to avoid exposing the process to the R session, and I think your functions are self-contained enough that their Python nature can be abstracted away like C(++). > > It is probably possible to fetch Python help pages in a safer way, if that is all that is required.
2020-04-03
Vince Carey (07:15:19) (in thread): > Yes, that safer way, probably through reticulate, sounds good. I think this version of BiocSklearn will be pretty conservative and always call basiliskEnd. Thanks for clarifying.
2020-05-02
Andrew Jaffe (15:47:17): > @Andrew Jaffe has joined the channel
2020-05-04
Nitin Sharma (06:27:37): > @Nitin Sharma has joined the channel
2020-06-06
Olagunju Abdulrahman (19:57:29): > @Olagunju Abdulrahman has joined the channel
2020-06-26
Jenny Smith (15:58:18): > @Jenny Smith has joined the channel
2020-07-07
Vivek Das (02:56:35): > @Vivek Das has joined the channel
2020-07-09
David Zhang (04:24:48): > @David Zhang has joined the channel
2020-07-15
wmuehlhaeuser (09:21:14): > @wmuehlhaeuser has joined the channel
2020-07-24
Isha Goel (07:48:44): > @Isha Goel has joined the channel
2020-07-29
Riyue Sunny Bao (17:37:45): > @Riyue Sunny Bao has joined the channel
2020-07-30
Ayush Raman (12:43:34): > @Ayush Raman has joined the channel
2020-07-31
Saulius Lukauskas (03:15:46): > @Saulius Lukauskas has joined the channel
bogdan tanasa (13:06:08): > @bogdan tanasa has joined the channel
Hector Climente (13:15:47): > @Hector Climente has joined the channel
2020-08-05
shr19818 (13:45:41): > @shr19818 has joined the channel
2020-08-06
Laurent Gatto (13:01:06): > @Laurent Gatto has joined the channel
2020-08-17
Roye Rozov (02:09:54): > @Roye Rozov has joined the channel
2020-08-18
Will Macnair (09:09:12): > @Will Macnair has joined the channel
Daniel Baker (09:13:30): > @Daniel Baker has joined the channel
2020-09-04
Goutham Atla (08:24:26): > @Goutham Atla has joined the channel
2020-09-06
Bob Policastro (08:36:04): > @Bob Policastro has joined the channel
Bob Policastro (08:36:43): > @Bob Policastro has left the channel
2020-10-19
Chris Chiu (23:30:17): > @Chris Chiu has joined the channel
2020-10-23
Rebecca Howard (08:17:18): > @Rebecca Howard has joined the channel
ImranF (11:43:29): > @ImranF has joined the channel
2020-10-26
Jeffrey O’Brien (14:03:51): > @Jeffrey O’Brien has joined the channel
Jeffrey O’Brien (14:03:56): > @Jeffrey O’Brien has left the channel
2020-11-04
Krutika (11:59:28): > @Krutika has joined the channel
2020-11-28
Rory Stark (12:38:13): > @Rory Stark has joined the channel
2020-12-02
Konstantinos Geles (Constantinos Yeles) (05:43:27): > @Konstantinos Geles (Constantinos Yeles) has joined the channel
2020-12-04
Arvind Mer (12:45:55): > @Arvind Mer has joined the channel
2020-12-07
Aoi Senju (21:19:37): > @Aoi Senju has joined the channel
2020-12-12
Huipeng Li (00:39:47): > @Huipeng Li has joined the channel
2020-12-14
Thomas Naake (08:56:03): > @Thomas Naake has joined the channel
2020-12-24
Jose Marie Antonio Miñoza (16:38:29): > @Jose Marie Antonio Miñoza has joined the channel
2021-01-01
Bernd (14:05:40): > @Bernd has joined the channel
2021-01-22
Annajiat Alim Rasel (15:44:21): > @Annajiat Alim Rasel has joined the channel
2021-02-07
Mikhael Manurung (11:09:44): > @Mikhael Manurung has left the channel
2021-02-15
Stefania Oliveto (17:32:43): > @Stefania Oliveto has joined the channel
2021-02-23
Wynn Cheung (10:32:43): > @Wynn Cheung has joined the channel
2021-02-24
Kyle Alford (11:40:43): > @Kyle Alford has joined the channel
2021-02-26
Dario Strbenac (21:00:08): > The random forest classifier was developed on small-dimensional, non-omics data sets in 2001. - File (PNG): image.png
Dario Strbenac (21:00:13): > Many bioinformatics journal articles and textbooks promote thesqrt(p)value for the number of variables to try at each split, often referred to asmtry. Say you have a gene expression dataset of 20000 genes and you filter lowly expressed genes and have 10000 genes remaining for classification.sqrt(p)is100. So, at each split, 9900 out of 10000 genes (99%) are ignored. It’s quite a different scenario tomtrybeing equal to 4 when the data set only has 16 variables, for example. What default value or range of trialled values do you use?
Aaron Lun (21:04:53): > Yes, I really struggled with random forests not picking the right features for cell type classification. Ended up having to tighten the search by giving it marker genes.
Dario Strbenac (21:15:04): > It’s interestnig you had to pre-select the features. Frank Harrell isagainst itbecause the classifier tends to find highly non-linear decision rules which no comparable feature selection algorithm would identify.
2021-03-03
Aedin Culhane (09:54:15): > Random forests (old style at least) used to hate too many features. For most other approaches SVM/NN you can get away without pre-select, though many do
Alan O’C (10:07:04): > Is this the case with regularised RFs?
Alan O’C (10:08:05): > Coincidentally just had an interesting journal club on DART, which seems a nice alternative esp for high-d data, if without the computational niceness of RF
Dario Strbenac (19:00:24) (in thread): > Did you adjustmtryfrom its default value which issqrt(p)by default and insuitable ifpis in the order of 20000.
2021-03-12
Julia Romanowska (14:51:33): > @Julia Romanowska has joined the channel
2021-03-19
Aedin Culhane (14:16:55): > @Alan O’CCan you send reference to DART or link to video if the presentation was recorded?>
Alan O’C (14:17:12): > It was not recorded
2021-03-23
Lambda Moses (23:05:21): > @Lambda Moses has joined the channel
2021-04-05
Maya Reed McDaniel (11:24:07): > @Maya Reed McDaniel has joined the channel
2021-04-21
Aishwarya Mandava (13:56:08): > @Aishwarya Mandava has joined the channel
2021-04-28
Mateusz Staniak (17:57:12): > @Mateusz Staniak has joined the channel
2021-05-01
Arjun Krishnan (21:43:09): > @Arjun Krishnan has joined the channel
2021-05-10
Jason Prasad (20:21:32): > @Jason Prasad has joined the channel
2021-05-11
Megha Lal (16:44:18): > @Megha Lal has joined the channel
2021-05-22
Moritz E. Beber (16:01:25): > @Moritz E. Beber has joined the channel
2021-05-25
Quang Nguyen (12:19:12): > @Quang Nguyen has joined the channel
2021-05-26
Federico Marini (03:02:51): > @Federico Marini has joined the channel
2021-06-04
Flavio Lombardo (05:51:53): > @Flavio Lombardo has joined the channel
2021-06-21
Andrea Gobbini (03:22:38): > @Andrea Gobbini has joined the channel
2021-06-24
Ilaria Billato (08:31:29): > @Ilaria Billato has joined the channel
2021-07-16
Nitesh Turaga (10:44:42): > @Nitesh Turaga has joined the channel
2021-07-19
Leo Lahti (17:01:39): > @Leo Lahti has joined the channel
2021-07-23
Batool Almarzouq (15:53:09): > @Batool Almarzouq has joined the channel
2021-07-26
Frank (20:18:02): > @Frank has joined the channel
2021-07-29
Gurpreet Kaur (01:03:11): > @Gurpreet Kaur has joined the channel
2021-08-04
TimNieuwenhuis (12:16:27): > @TimNieuwenhuis has joined the channel
2021-08-06
ChiaSin (20:01:47): > @ChiaSin has joined the channel
2021-08-10
Woo (08:18:02): > @Woo has joined the channel
2021-08-19
Ava Hoffman (she/her) (11:33:30): > @Ava Hoffman (she/her) has joined the channel
2021-09-01
Katie Saund (14:43:12): > @Katie Saund has joined the channel
2021-09-06
Eddie (08:22:54): > @Eddie has joined the channel
Eddie (09:09:06): > @Eddie has left the channel
2021-09-08
Espen Riskedal (06:10:36): > @Espen Riskedal has joined the channel
2021-09-16
Henry Miller (18:36:01): > @Henry Miller has joined the channel
2021-09-25
Mikey C (19:05:27): > @Mikey C has joined the channel
2021-10-27
Nicholas Cooley (11:02:18): > @Nicholas Cooley has joined the channel
2021-11-08
Paula Nieto García (03:28:31): > @Paula Nieto García has joined the channel
2022-01-03
Kurt Showmaker (17:04:39): > @Kurt Showmaker has joined the channel
2022-02-15
Gene Cutler (12:00:51): > @Gene Cutler has joined the channel
2022-02-23
Paulina Paiz (10:45:42): > @Paulina Paiz has joined the channel
Susanna (12:08:26): > @Susanna has joined the channel
2022-03-05
Anshul Kundaje (20:26:40): > @Anshul Kundaje has joined the channel
2022-03-15
Jeanette Johnson (12:04:10): > @Jeanette Johnson has joined the channel
2022-03-20
Sarvesh Nikumbh (21:41:46): > @Sarvesh Nikumbh has joined the channel
2022-03-21
Pedro Sanchez (05:01:37): > @Pedro Sanchez has joined the channel
2022-04-12
Vivian Chu (13:59:31): > @Vivian Chu has joined the channel
2022-05-03
Ray Su (06:56:00): > @Ray Su has joined the channel
2022-05-20
Simon Pearce (03:16:06): > @Simon Pearce has joined the channel
2022-06-09
John Hutchinson (09:08:18): > @John Hutchinson has joined the channel
2022-06-12
Karat Sidhu (12:37:27): > @Karat Sidhu has joined the channel
2022-07-13
Indrik Wijaya (21:27:47): > @Indrik Wijaya has joined the channel
2022-07-15
Ashley Robbins (15:23:52): > @Ashley Robbins has joined the channel
2022-07-18
Filip Stachura (09:32:53): > @Filip Stachura has joined the channel
2022-07-28
Krithika Bhuvanesh (13:52:38): > @Krithika Bhuvanesh has joined the channel
2022-08-11
Rene Welch (17:15:27): > @Rene Welch has joined the channel
2022-08-30
Lisa Breckels (07:20:49): > @Lisa Breckels has joined the channel
2022-09-27
Jennifer Holmes (16:15:18): > @Jennifer Holmes has joined the channel
2022-10-06
Devika Agarwal (05:37:20): > @Devika Agarwal has joined the channel
2022-10-10
Mercilena Benjamin (13:56:11): > @Mercilena Benjamin has joined the channel
2022-10-31
Chenyue Lu (10:05:04): > @Chenyue Lu has joined the channel
2022-11-06
Sherine Khalafalla Saber (11:20:53): > @Sherine Khalafalla Saber has joined the channel
2022-12-01
Chris Chiu (05:23:51): > @Chris Chiu has left the channel
2022-12-02
Atuhurira Kirabo Kakopo (15:59:33): > @Atuhurira Kirabo Kakopo has joined the channel
2022-12-12
Zainab Ashimiyu-Abdusalam (04:53:33): > @Zainab Ashimiyu-Abdusalam has joined the channel
Lexi Bounds (17:58:57): > @Lexi Bounds has joined the channel
2022-12-13
Ana Cristina Guerra de Souza (09:00:18): > @Ana Cristina Guerra de Souza has joined the channel
Xiangnan Xu (18:32:11): > @Xiangnan Xu has joined the channel
Helen Fu (21:53:53): > @Helen Fu has joined the channel
2022-12-14
Lijia Yu (19:38:56): > @Lijia Yu has joined the channel
2022-12-20
Krithika Bhuvanesh (13:24:45): > @Krithika Bhuvanesh has left the channel
2023-01-10
Vince Carey (10:49:31): > @Vince Carey has left the channel
2023-01-21
Hien (16:02:53): > @Hien has joined the channel
2023-01-31
Ahmad Al Ajami (09:10:16): > @Ahmad Al Ajami has joined the channel
Ahmad Al Ajami (09:10:19): > @Ahmad Al Ajami has left the channel
2023-02-09
Wes W (22:03:47): > @Wes W has joined the channel
2023-02-22
michaelkleymn (01:50:55): > @michaelkleymn has joined the channel
2023-02-28
Ramin (15:31:18): > @Ramin has joined the channel
Krithika Bhuvanesh (16:53:49): > @Krithika Bhuvanesh has joined the channel
2023-03-01
jeremymchacón (12:13:02): > @jeremymchacón has joined the channel
2023-03-07
Thomas Klammsteiner (08:35:33): > @Thomas Klammsteiner has joined the channel
2023-03-10
Joaquin Reyna (15:35:51): > @Joaquin Reyna has joined the channel
2023-03-22
Keerthana (08:08:32): > @Keerthana has joined the channel
2023-05-04
Leopoldo Valiente (16:27:17): > @Leopoldo Valiente has joined the channel
2023-05-17
Hassan Kehinde Ajulo (12:17:40): > @Hassan Kehinde Ajulo has joined the channel
2023-05-18
Oluwafemi Oyedele (05:54:44): > @Oluwafemi Oyedele has joined the channel
2023-05-31
Alyssa Obermayer (14:15:25): > @Alyssa Obermayer has joined the channel
2023-06-07
Alyssa Obermayer (18:29:53): > @Alyssa Obermayer has joined the channel
2023-06-19
Pierre-Paul Axisa (05:11:24): > @Pierre-Paul Axisa has joined the channel
2023-07-13
Brian Schilder (07:02:26): > @Brian Schilder has joined the channel
2023-07-28
Benjamin Yang (15:57:51): > @Benjamin Yang has joined the channel
2023-08-02
Beth Cimini (08:20:54): > @Beth Cimini has joined the channel
2023-08-03
Ritika Giri (15:58:23): > @Ritika Giri has joined the channel
2023-08-04
Trisha Timpug (09:35:41): > @Trisha Timpug has joined the channel
2023-08-18
Victor Yuan (12:35:57): > @Victor Yuan has joined the channel
2023-08-20
Jacques SERIZAY (10:38:40): > @Jacques SERIZAY has joined the channel
2023-09-03
Lea Seep (09:51:45): > @Lea Seep has joined the channel
2023-09-15
Leo Lahti (04:55:17): > @Leo Lahti has joined the channel
2023-10-11
Nikhil Mane (13:59:37): > @Nikhil Mane has joined the channel
2023-11-16
Sue McClatchy (10:00:19): > @Sue McClatchy has joined the channel
2023-11-18
karisssq (22:33:22): > @karisssq has joined the channel
2023-11-21
Konstantinos Geles (Constantinos Yeles) (05:46:46): > @Konstantinos Geles (Constantinos Yeles) has left the channel
2023-12-01
Tram Nguyen (10:16:58): > @Tram Nguyen has joined the channel
2023-12-04
Ajda Pristavec (12:08:55): > @Ajda Pristavec has joined the channel
2024-01-11
Nilesh Kumar (12:01:45): > @Nilesh Kumar has joined the channel
2024-01-26
Brian Schilder (09:32:08): > For those who are interested in making ML model withpytorch, but want to stay within the R ecosystem, I highly recommend thetorchpackage! They provide an excellent tutorial that explains principles of torch and ML more generally in a very accessible way:https://skeydan.github.io/Deep-Learning-and-Scientific-Computing-with-R-torch/
2024-01-27
atongsa miyamoto (04:52:51): > @atongsa miyamoto has joined the channel
2024-02-16
Jovana Maksimovic (03:00:24): > @Jovana Maksimovic has joined the channel
Vince Carey (06:40:09): > @Vince Carey has joined the channel
Vince Carey (06:52:49): > Thanks to support from NHGRI, Bioconductor now has an NVIDIA L40S GPU to use for development and testing. A basic objective is to add regular checking and evaluation of Bioc packages that have code that can take advantage of GPU availability. It would be great to develop some documentation on effective practices for doing modern machine learning in genomics. For instance, there are various “ways in” to using torch in R: torch, luz, brulee, … How does one navigate in this area?
Vince Carey (06:54:34): > We’d also like to improve concepts of instrumentation in this area – how do you measure efficiency of use of GPU and CPU and how can we code to ensure we hit reasonable efficiency objectives?
Maria Doyle (09:05:53): > @Maria Doyle has joined the channel
Dirk Eddelbuettel (09:32:35): > @Dirk Eddelbuettel has joined the channel
António Domingues (10:50:50): > @António Domingues has joined the channel
Hervé Pagès (11:53:35): > @Hervé Pagès has joined the channel
jim rothstein (12:24:38): > @jim rothstein has joined the channel
2024-02-19
Alex Bott (17:24:42): > @Alex Bott has joined the channel
2024-02-28
Louise Morlot (10:48:02): > @Louise Morlot has joined the channel
2024-03-10
Cherishma Subhasa (19:57:48): > @Cherishma Subhasa has joined the channel
2024-03-27
abhich (05:45:19): > @abhich has joined the channel
2024-04-04
Alexandru Mizeranschi (09:37:11): > @Alexandru Mizeranschi has joined the channel
Tung Trinh (23:39:07): > @Tung Trinh has joined the channel
2024-04-18
Philipp Sergeev (03:02:21): > @Philipp Sergeev has joined the channel
Jianhai Zhang (18:28:52): > @Jianhai Zhang has joined the channel
2024-04-28
Danielle Callan (08:32:38): > @Danielle Callan has joined the channel
2024-05-05
Aedin Culhane (03:35:58): > Should this be relabelled AI-ML??? Just a thought.
Aedin Culhane (03:36:01): > CZI Open Science, in collaboration with Sloan Foundation, has opened an RFI focused on AI-enabled tools for scientific research. > > “Are you developing tools for #AI enabled research practices that also promote open research? We want to hear from you! Submit your ideas to our RFI so we can build a more effective scientific community together.https://bit.ly/49ZUR26” - Attachment (Medium): Request for Information: Pathways to AI-enabled research > Transforming scientific work with tools and technology
2024-05-06
Michal Kolář (11:57:14): > @Michal Kolář has joined the channel
2024-05-11
Aleru Divine (04:11:24): > @Aleru Divine has joined the channel
2024-05-14
Claire Seibold (11:48:30): > @Claire Seibold has joined the channel
2024-06-11
Ziru Chen (04:37:04): > @Ziru Chen has joined the channel
2024-07-02
Diána Pejtsik (10:59:59): > @Diána Pejtsik has joined the channel
2024-07-05
Antonin Thiébaut (04:37:56): > @Antonin Thiébaut has joined the channel
2024-07-11
Hothri Moka (07:20:54): > @Hothri Moka has joined the channel
2024-07-15
Erick Navarro (16:20:05): > @Erick Navarro has joined the channel
2024-07-26
Qiwen Octavia Huang (19:51:40): > @Qiwen Octavia Huang has joined the channel
2024-07-29
JP Flores (17:09:16): > @JP Flores has joined the channel
2024-07-31
Zahraa W Alsafwani (17:24:29): > @Zahraa W Alsafwani has joined the channel
2024-08-04
sunnyday (14:58:27): > @sunnyday has joined the channel
2024-08-19
Rema Gesaka (09:41:52): > @Rema Gesaka has joined the channel
2024-09-10
Alex Qin (03:49:20): > @Alex Qin has joined the channel
2024-10-01
Caroline Schreiber (04:10:02): > @Caroline Schreiber has joined the channel
2024-10-12
Francisco Ignacio Oñate Vidal (00:50:35): > @Francisco Ignacio Oñate Vidal has joined the channel
2024-10-23
Sounkou Mahamane Toure (11:49:20): > @Sounkou Mahamane Toure has joined the channel
2024-11-11
Josh Steier (12:42:30): > @Josh Steier has joined the channel
Josh Steier (13:03:45): > Hi everyone, > I have a background in computational biology(MS applied math and stats degree from Stony Brook University), and a strong background in machine learning, where I currently work as a machine learning engineer. > I’ve been working in machine learning for roughly 3-4 years, including my educational background, and programming in Python since I was younger. > I’m eager to contribute to Bioconductor and don’t know where to start, can someone help me with that? > Thank you!
Mikhail Dozmorov (13:06:50): > @Mikhail Dozmorov has joined the channel
2024-11-12
Kylie Bemis (09:21:30): > @Kylie Bemis has joined the channel
Marouen Ben Guebila (22:33:59): > @Marouen Ben Guebila has joined the channel
2024-11-21
Nesrine Gharbi (04:12:33): > @Nesrine Gharbi has joined the channel
2025-03-18
Andres Wokaty (14:27:59): > @Andres Wokaty has joined the channel
Nicolo (15:00:27): > @Nicolo has joined the channel
2025-04-14
Saad Farooq (07:09:41): > @Saad Farooq has joined the channel
2025-04-15
Rasmus Hindström (07:08:27): > @Rasmus Hindström has joined the channel
2025-04-17
dhvani solanki (16:17:42): > @dhvani solanki has joined the channel
2025-04-18
Juan Henao (04:43:15): > @Juan Henao has joined the channel