James D Wilson

@sp_monte_carlo It is an interesting but important change of perspective. Developing generative models for weighted graphs can be challenging too. I’ve worked some in this space for weighted edge weights and correlation “networks.” See sciencedirect.com/science/articl… nature.com/articles/s4159…

@TattedMiscreant I’ve heard the dry erase marker over top trick but never tried it. But I do know that acetone (primary agent in finger nail polish remover) will take it right off

I’d like to use my skills as a statistician, network scientist, and ML methodologist to help tackle big environmental and climate issues. Does anyone have good references or problems to look into?

Really interesting mathematical result on transformers. Nice work

@A_Kolobaric Congratulations Dr. Kolobaric!

Excited to enter 2024 with my PhD!

@francescadomin8 @HarvardBigData @harvard_data @WiDS_Worldwide @Conf_CLeaR @AICLIMATE Hi @francescadomin8 I am very interested in learning more about what you are working on. I am a Ph.D statistician at @usfca and wanting to help where I can with climate research. If you are also looking for possible research collaborations please let me know.

Interested in #datascience, #causalinference, #climate? We have several job opportunities in my lab. They are all listed here: hsph.harvard.edu/nsaph/opportun… including post-docs, data scientists. Join us! @HarvardBigData @harvard_data @WiDS_Worldwide @Conf_CLeaR @AICLIMATE

New paper in Alzheimer's & Dementia🚨 We: 1) introduce a new method for measuring in vivo #whitematterinjury, unhealthy white matter connectivity (UWMC) 2) quantify UWMC in brain regions affected by AB+tau 3) compare UWMC between racialized groups Link: alz-journals.onlinelibrary.wiley.com/doi/10.1002/al…

@SrijanSengupta7 For DMS, it should just really be called “Applications”

The existence of broader impacts implies the existence of narrower impacts. Or just regular impacts.

@tiagopeixoto This brings me back to a key question of when, in general, a graph (and hence graph tools/analyses) is needed/useful for observed data. Of course some data are inherently graph-based (friendships, followers, etc.) but for those that are not, I feel a statistical test is needed

It was always clear that graph neural networks don't actually need the graph for most tasks and don't perform so well when the node features are not doing the heavy lifting. But it seems that often including the graph makes things *worse*: arxiv.org/abs/2309.04332

@AGerlachPhD Congratulations Andrew! I’m really happy for you and am excited to see what you do next!

Difficult to describe how proud I am to share this! I’ve put in a LOT of hard work to get here, but I’ve also had tremendous support along the way from family, friends, colleagues, and mentors. But mostly, I just wish my mom were still here to see this.

@AGerlachPhD Huge congratulations Andrew! Well deserved! And what great timing

Can’t get much better than a week in Yosemite Valley with the family… unless you get your K01 notice of grant award while there!!!

Check out our new paper investigating sex differences in functional connectivity response to antidepressant treatment for those in late-life. The findings here may greatly reduce the time needed to assess pharmacological treatment efficacy for those with late-life depression

@roydanroy Anyway, some cool work is being done in the statistical network analysis area. See this as a fun example where the focus is on identifying anomalous subgraphs: arxiv.org/pdf/1807.08925…

@roydanroy I’m used to thinking about anomalies as data points (or vectors / networks) that are outside of a predictive interval for a new value given “normative” data. So something like anything outside 3 standard deviations from the mean.

Who's an expert in anomaly detection out there? What's are the key texts in this area? Does it have a rigorous theoretical foundation? That handles the adversarial aspects? And what about deep learning and in-roads there?

I’m so proud of our new work led by Elisabeth Salisbury of @PittPsychiatry. Personally as a dad of 3, I’m particularly excited about the potential for helping newborns in need. I am looking forward to our next work, and also I hope to see these findings put into practice.

Excited to share that our work lead by Ahmad Mayeli on alterations in RAR and negative symptoms in SSD was published today on @molpsychiatry. Thanks to @AliceLaGoy, @ThisIsJDWilson, @armandagostino, and the DIAPASON group. For more info, here is the link: rdcu.be/c9T67

@FabFerrare @PittPsychiatry @PittTweet Congratulations! Well deserved Fabio!

Tenured. Thank you @PittPsychiatry, @PittTweet. Looking forward to what is coming next.

Excited to see some of our recent work mentioned in a great piece on @NBCNightlyNews about predicting political leaning with fMRI (see time 1:43). Thank you for the mention @byjacobward . I look forward to seeing where AI and brain imaging goes from here.

@cgibbs_10 @baileyfosdick @rtelmore @KoslovskyMatt @andeekaplan @JuliaSharpStats @ryanlayer @michael_boulder It has been a lot of fun Connor! I’m excited to see what you do next

@baileyfosdick @rtelmore @ThisIsJDWilson @KoslovskyMatt @andeekaplan @JuliaSharpStats @ryanlayer @michael_boulder, thank you for all of your help and support! Also, a shout out to Brooke Anderson, Felix Duerr, and Sky Martin for their help as well.

I'm excited to share that on March 6, I successfully defended my dissertation, "Causality and Clustering in Complex Settings," and I am now officially a PhD in Statistics! I'm excited to see what comes next.