A
Arvind Neelakantan
Researcher at University of Massachusetts Amherst
Publications - 41
Citations - 15055
Arvind Neelakantan is an academic researcher from University of Massachusetts Amherst. The author has contributed to research in topics: Artificial neural network & Knowledge base. The author has an hindex of 23, co-authored 37 publications receiving 5594 citations. Previous affiliations of Arvind Neelakantan include BBN Technologies & Google.
Papers
More filters
Towards a better understanding of Vector Quantized Autoencoders
TL;DR: This work investigates an alternate training technique for VQ-VAE, inspired by its connection to the Expectation Maximization (EM) algorithm, and develops a non-autoregressive machine translation model whose accuracy almost matches a strong greedy autoregressive baseline Transformer, while being 3.3 times faster at inference.
Proceedings Article
RelNet: End-to-end Modeling of Entities & Relations.
TL;DR: RelNet as mentioned in this paper is a memory augmented neural network which models entities as abstract memory slots and is equipped with an additional relational memory which models relations between all memory pairs, thus building an abstract knowledge graph on the entities and relations present in a document which can be used to answer questions about the document.
Posted Content
Efficient Non-parametric Estimation of Multiple Embeddings per Word in Vector Space
TL;DR: This paper proposed an extension to the Skip-gram model that efficiently learns multiple embeddings per word type, which differs from recent related work by jointly performing word sense discrimination and embedding learning, by non-parametrically estimating the number of senses per word and by its efficiency and scalability.
Active Learning is a Strong Baseline for Data Subset Selection
Tom B. Brown,N. C. Ryder,Jared Kaplan,Prafulla Dhariwal,Arvind Neelakantan,Pranav Shyam,Girish Sastry,Alec Radford,Chris Hallacy,Aditya Ramesh,Gabriel Goh,Sandhini Agarwal,Alexey Dosovitskiy,Lucas Beyer,Alexander Kolesnikov,Dirk Weissenborn,Xiaohua Zhai,Sören Mindermann,Jan Markus Brauner +18 more
TL;DR: A simple active learning-based algorithm that outperforms all the current data subset selection algorithms on the benchmark tasks and finds that it is crucial to find a balance between easy- to-classify and hard-to- classify examples when selecting a subset.
Knowledge Representation and Reasoning with Deep Neural Networks
TL;DR: This research presents a novel and scalable approach called “Smart grids” that combines nanofiltration and “smart cities” to solve the challenge of integrating smart phones and smart grids into the offline world.