Eptehal Nashnoush

Research Engineer | Teaching Fellow @MIT

[neuroccm @neuroccm]

Congratulations to @EptehalN @hamzamahdi97 @rishit_dagli @rami_saab @ArjBalachandar🙌, presenting our work on machine learning to detect swallowing difficulty using speech audio as a biomarker, using prospectively collected small datasets at #CHIL2024 #ML #AI4QI (AI for quality improvement in clinical care); thanks to @UofT_TCAIREM @uoftmedicine @Sunnybrook for support

[chilconference @CHILconference]

Learn about: "Tuning In : Analysis of Audio Classifier Performance in Clinical Settings with Limited Data" by @rami_saab @EptehalN @hamzamahdi97 @ArjBalachandar @rishit_dagli @lucas_xperri @neuroccm at #CHIL2024!

[Eptehal Nashnoush @EptehalN]

@neuroccm @ArjBalachandar @uoftneurons Congrats @ArjBalachandar

[neuroccm @neuroccm]

🌊🔊Exciting work from: our stroke innovation lab, project MASA, we compare deep learning models like CNNs (DenseNet, ConvNeXt) & transformers (ViT, SWIN, AST) for audio classification on small clinical datasets. Pre-training on large datasets before fine-tuning & careful spectrogram preprocessing are key for strong performance, as tested on first-of-their-kind prospectively collected stroke patient audio. Our findings highlight the importance of model selection, pre-training & preprocessing for audio-based clinical diagnostics - @hamzamahdi97 @EptehalN @rami_saab @ArjBalachandar @rishit_dagli #CHIL2024

[neuroccm @neuroccm]

Delighted to share that our #AI4QI manuscript has been accepted at #CHIL2024 and will be published, super kudos! @hamzamahdi97 @EptehalN @rami_saab @ArjBalachandar @rishit_dagli Tuning In: Analysis of Audio Classifier Performance in Clinical Settings with Limited Data🔊🌊, we evaluated the performance of various deep learning models for audio classification in clinical settings, focusing on small datasets that reflect real-world prospective data collection and its challenges. We analyze CNNs, transformer models, and pre-trained audio models, highlighting the benefits of pre-training on large datasets before fine-tuning on specific clinical data. To conduct our research, we collected two first-of-their-kind patient audio datasets from #stroke patients and investigated various preprocessing techniques. Our findings indicate that CNNs can match or exceed transformer models in small dataset contexts, with DenseNet-Contrastive and AST models showing notable performance. We underscore the significance of incremental marginal gains through model selection, pre-training, and preprocessing in sound classification. We believe that our study offers valuable insights for clinical diagnostics that rely on audio classification.
Thanks to @sunnybrook AFP, @UofT_TCAIREM , @uoftmedicine @UofTNeurology for their generous support of our work in #ML #AI in quality improvement for stroke and other neurology patients. arXiv version: https://t.co/Z7KookQF8k; We thank all 4 reviewers for their valuable and insightful comments.

[neuroccm @neuroccm]

Delighted to share work from our lab on using audio as a biomarker in clinical settings, in this case expanding on its use for classifying swallowing impairments with limited datasets #ML4H #ML4QI #Stroke https://t.co/lf1twalPHI

[neuroccm @neuroccm]

@hamzamahdi97 @EptehalN @rami_saab @ArjBalachandar @rishit_dagli
Blog post: https://t.co/fACFzzaV9S

[Temerty Centre for AI in Medicine (T-CAIREM) @UofT_TCAIREM]

📣 March 2024 @UofT_TCAIREM #newsletter is out!
🏆 New DSI/T-CAIREM grant winners announced
🎥 VIDEO: @kdpsinghlab “Why AI Health Fails”
🎥 VIDEO: #Toronto #Health #Datathon Panel Discussion with @EptehalN, @alistairewj, @obenfine, and Micaela Consens
🔗 https://t.co/yADiGtFvtB

[Vector Institute @VectorInst]

🚨⏰ Hurry and secure your spot for tomorrow’s webinar! Learn if you’re eligible for Vector’s 17.5K Scholarship in AI and meet past scholarship recipients @EptehalN + @anthony_wong655 for a fireside chat!

https://t.co/7norx7Ouo0

[neuroccm @neuroccm]

Incredibly proud of @EptehalN for her epic presentation @UofT_TCAIREM as part of our lab @hamzamahdi97 @rami_saab @ArjBalachandar 💥🎉@NQILTeam #AI4QI ML for quality improvement

[Nova Scotia Gov. @nsgov]

Nova Scotia is announcing no new cases of COVID-19 today, June 23. There are a total of 60 active cases in the province.

For more, read the news release: https://t.co/JQ4eZjYe4U

[McMaster Health Sciences @machealthsci]

. @machealthsci’s Ishac Nazy has discovered what caused potentially deadly #clots in #COVID19 patients. https://t.co/hMlcX20iL7 #McMasterUniversity #Hamilton #pandemic

[GIT Coatings @gitcoatings]

We are proud to be one of the companies the spun out of @DalhousieU.

The @dalideahub was there every step of the way. Thank you!

[Liz (Welsh) Beechinor 😅 @LizTweetsTech]

My dad led an organization of over 300,000 employees.

When I took on my biggest marketing & events team to date (35 employees) I asked him how the *heck* he did it. Here's what he said and what helped me every day. 🧵

[Dal Grad Studies @DalGradStudies]

How are you supposed to talk to a customer? This workshop teaches you, a future entrepreneur, how to do it! via @DalInnovates

Thu, March 25, 2021
12:00 PM – 1:30 PM ADT

https://t.co/49dSjxxuLY

[Atlantic Student Research Journal @TheASRJ]

The ASRJ is excited to present 'ENGspire 2021', an event of the spire series. Register for this webinar using the link below:
https://t.co/YkMj9LL0bf
.
Feel free to reach out if you have any questions!
.
#TheASRJ #spireseries #ENGspire #engineering #research #conference

[@DALideaHUB @dalideahub]

Our workshop on "Legal Fundamentals" with Jacqui Walsh is starting at 10:00 a.m.! Join us for this 2-hour session in partnership with @voltaeffect and have all your questions answered. We'll see you there! https://t.co/zf3zmSAsP9

[@DALideaHUB @dalideahub]

Our #HUBNerds have nothing but good things to say about Jacqui and the Initio Law Clinic.

Jacqui and her team are presenting a workshop on "Legal Fundamentals" next Tuesday, and we want to extend an invitation to our entire community.

Register here: https://t.co/3ws6ZSyEi7