Multitask Prompted Training Enables Zero-Shot Task Generalization
- Victor Sanh, Albert Webson, Alexander M. Rush
- Computer ScienceInternational Conference on Learning…
- 15 October 2021
A system for easily mapping any natural language tasks into a human-readable prompted form and fine-tune a pretrained encoder-decoder model on this multitask mixture covering a wide variety of tasks.
Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models
- A. Srivastava, Abhinav Rastogi, Uri Shaham
- Computer ScienceArXiv
- 9 June 2022
Evaluation of OpenAI's GPT models, Google-internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters finds that model performance and calibration both improve with scale, but are poor in absolute terms.
BLOOM: A 176B-Parameter Open-Access Multilingual Language Model
- Teven Le Scao, Angela Fan, Thomas Wolf
- Computer ScienceArXiv
- 9 November 2022
BLOOM is a 176B-parameter open-access language model designed and built thanks to a collaboration of hundreds of researchers and achieves competitive performance on a wide variety of benchmarks, with stronger results after undergoing multitask prompted finetuning.
Caspase-1 causes truncation and aggregation of the Parkinson’s disease-associated protein α-synuclein
- Wei Wang, L. Nguyen, Q. Q. Hoang
- Biology, ChemistryProceedings of the National Academy of Sciences
- 1 August 2016
It is reported that inflammation can play a role in causing the aggregation of wild type aSyn, and perhaps in the pathogenesis of sporadic PD as well.
An allosteric circuit in caspase-1.
- Debajyoti Datta, J. Scheer, M. J. Romanowski, J. Wells
- Chemistry, BiologyJournal of Molecular Biology
- 19 September 2008
Substrate and Inhibitor-induced Dimerization and Cooperativity in Caspase-1 but Not Caspase-3*
- Debajyoti Datta, Christopher L. McClendon, M. Jacobson, J. Wells
- Biology, ChemistryJournal of Biological Chemistry
- 5 February 2013
Steady-state kinetic analysis and modeling independently support this model, where binding of one substrate molecule not only increases substrate binding in preformed dimers but also drives the formation of heterodimers.
learn2learn: A Library for Meta-Learning Research
- Sébastien M. R. Arnold, Praateek Mahajan, Debajyoti Datta, Ian Bunner, Konstantinos Saitas Zarkias
- Computer ScienceArXiv
- 27 August 2020
Meta-learning researchers face two fundamental issues in their empirical work: prototyping and reproducibility. Researchers are prone to make mistakes when prototyping new algorithms and tasks…
PatientExploreR: an extensible application for dynamic visualization of patient clinical history from electronic health records in the OMOP common data model
- B. Glicksberg, B. Oskotsky, A. Butte
- Computer ScienceBioinform.
- 19 June 2019
PatientExploreR is an extensible application built on the R/Shiny framework that interfaces with a relational database of EHR data in the Observational Medical Outcomes Partnership CDM format and produces patient-level interactive and dynamic reports and facilitates visualization of clinical data without any programming required.
Synaptotagmin perturbs the structure of phospholipid bilayers.
- V. Shahin, Debajyoti Datta, E. Hui, R. Henderson, E. Chapman, J. Edwardson
- Biology, ChemistryBiochemistry
- 19 February 2008
It is proposed that bilayer perturbation by syt might be significant with respect to its ability to promote membrane fusion in response to Ca2+ binding.
Influenza-like symptom recognition using mobile sensing and graph neural networks
- Guimin Dong, Lihua Cai, Debajyoti Datta, Shashwat Kumar, Laura E. Barnes, M. Boukhechba
- Computer ScienceACM Conference on Health, Inference, and Learning
- 8 April 2021
This work proposes an end-to-end graph neural network model with multi-channel mobile sensing input for influenzalike symptom recognition based on people's daily mobility, social interactions, and physical activities and shows that GNN with GraphSAGE convolutional layers significantly outperforms baseline models with handcrafted features.
...
...