Samira Abnar

Computer Science · Physical Sciences

11h-index636citations33works0.22yr avg
OpenAlex↗Google Scholar↗

Accepting Students?

No reports yet. Know if this professor is taking students?

Research Topics

Topic Modeling(17), Natural Language Processing Techniques(10), Multimodal Machine Learning Applications(7), Domain Adaptation and Few-Shot Learning(6), Neurobiology of Language and Bilingualism(5)

Publications33 total

How PARTs assemble into wholes: Learning the relative composition of images
arXiv (Cornell University)·2025Open Access
Parameters vs FLOPs: Scaling Laws for Optimal Sparsity for Mixture-of-Experts Language Models
arXiv (Cornell University)·2025Open Access
How Parts Assemble into Wholes:Learning the Relative Composition of Images
SSRN Electronic Journal·2025Open Access
Aggregate-and-Adapt Natural Language Prompts for Downstream Generalization of CLIP
arXiv (Cornell University)·2024· 1 citedOpen Access
Aggregate-and-Adapt Natural Language Prompts for Downstream Generalization of CLIP
2024
Adaptivity and Modularity for Efficient Generalization Over Task Complexity
arXiv (Cornell University)·2023Open Access
Diffusion Probabilistic Fields
arXiv (Cornell University)·2023· 1 citedOpen Access
Robust Evaluation of Language–Brain Encoding Experiments
Lecture notes in computer science·2023· 5 citedOpen Access
Scaling Laws vs Model Architectures: How does Inductive Bias Influence Scaling?
2023· 27 citedOpen Access
GAUDI: A Neural Architect for Immersive 3D Scene Generation
arXiv (Cornell University)·2022· 55 citedOpen Access
Scaling Laws vs Model Architectures: How does Inductive Bias Influence Scaling?
arXiv (Cornell University)·2022· 11 citedOpen Access
Exploring the Limits of Large Scale Pre-training
arXiv (Cornell University)·2021· 35 citedOpen Access
Scale Efficiently: Insights from Pre-training and Fine-tuning Transformers
arXiv (Cornell University)·2021· 58 citedOpen Access
Gradual Domain Adaptation in the Wild:When Intermediate Distributions are Absent
arXiv (Cornell University)·2021· 4 citedOpen Access
Long Range Arena : A Benchmark for Efficient Transformers
2021· 57 cited
Long Range Arena: A Benchmark for Efficient Transformers
arXiv (Cornell University)·2020· 194 citedOpen Access
Transferring Inductive Biases through Knowledge Distillation
arXiv (Cornell University)·2020· 12 citedOpen Access
A Comparison of Architectures and Pretraining Methods for Contextualized Multilingual Word Embeddings
Proceedings of the AAAI Conference on Artificial Intelligence·2020· 7 citedOpen Access
Quantifying Attention Flow in Transformers
2020· 21 citedOpen Access
A Comparison of Architectures and Pretraining Methods for Contextualized\n Multilingual Word Embeddings
arXiv (Cornell University)·2019Open Access
page 1 of 2Next →

Frequent Co-authors

Mostafa Dehghani(12), Willem Zuidema(7), Josh Susskind(6), Yi Tay(5), Jinfeng Rao(5), Donald Metzler(5), Lisa Beinborn(5), Rochelle Choenni(5), Chen Huang(5), Vimal Thilak(4), William Fedus(3), Hyung Won Chung(3), Sharan Narang(3), Dani Yogatama(3), Miguel Ángel Bautista(3), Shuangfei Zhai(3), Hanlin Goh(3), Azadeh Shakery(3), Dan Busbridge(3), Yikang Shen(2)