From Group to Individual Labels Using Deep Features

By Dimitrios Kotzias, Misha Denil, Nando de Freitas, Padhraic Smyth. 2015

330 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning, Evaluation & Benchmarks, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

In Evaluation & Benchmarks

In Signal Processing

Citations in the corpus

The Natural Language Decathlon: Multitask Learning as Question Answering (2018, methodology): “On Amazon and Yelp reviews [Kotzias et al., 2015], a MQAN pretrained on decaNLP achieves exact match scores of 82.1% and 80.8%, respectively, without any fine-tuning.”
Finetuned Language Models Are Zero-Shot Learners (2021, background): “For instance, McCann et al. (2018) cast ten NLP tasks as QA and train a model on a collection of tasks formulated with natural language prompts; they report transfer learning gains on finetuning tasks as well as zero-shot domain adaptation results on SNLI (Bowman et al., 2015) and Amazon/Yelp Review”
Semantically Equivalent Adversarial Rules for Debugging NLP models (2018, methodology): “, 2016) model with unigrams and bigrams (embedding size of 50) on RottenTomato movie reviews (Pang and Lee, 2005), and evaluate it on IMDB sentence-sized reviews (Kotzias et al., 2015), simulating the common case where a model trained on a public dataset is applied to new data from a similar domain.”
Label-Free Supervision of Neural Networks with Physics and Domain Knowledge (2016, methodology): “(Kotzias et al. 2015) used constraint learning to train deep networks in a natural language setting.”
Deep Multi-instance Networks with Sparse Label Assignment for Whole Mammogram Classification (2016, methodology): “incorporated CNN features into multi-instance cost function to do sentiment analysis [12].”
Aspect-Controllable Opinion Summarization (2021, background): “The goal is then to infer labels for bags (Dietterich et al., 1997; Maron and Ratan, 1998) or jointly for instances and bags (Zhou et al., 2009; Wei et al., 2014; Kotzias et al., 2015; Xu and Lapata, 2019; Angelidis and Lapata, 2018a).”
Towards a Unified Natural Language Inference Framework to Evaluate Sentence Representations (2018, neutral): “…Inference MegaAttitude (White and Rawlins, 2016, 2018) 11K (11,814) Relationship Extraction FACC1 (Gabrilovich et al., 2013) 30K Sentiment Analysis (Kotzias et al., 2015) 6K (Yang et al., 2015) 7K (7,706) Puns SemEval 2017 Task 7 (Miller et al., 2017) 8K (8,054) Combined Unified NLI Framework 433K…”
A survey of neural network techniques in market trend analysis (2017, background): “Through multiple extractions, deep learning can obtain higher dimensional features for classification[14].”
Collecting Diverse Natural Language Inference Problems for Sentence Representation Evaluation (2018, methodology): “Phenomena Dataset # pairs Automated Decomp (Rudinger et al., 2018b) 42K (41,888) 3 UW (Lee et al., 2015) 5K (5,094) 3Event Factuality MeanTime (Minard et al., 2016) .7K (738) 3 Groningen (Bos et al., 2017) 260K (261,406) 3Named Entity Recognition CoNLL (Tjong Kim Sang and De Meulder, 2003) 60K (59,9”
Multiple instance learning: A survey of problem characteristics and applications (2016, methodology): “In [40, 125, 129, 130], MIL was used to infer the sentiment expressed in individual sentences based on the labels provided for entire user reviews.”
Weakly supervised learning via statistical sufficiency (2016, background): “Similarly, in sentiment analysis positive/negative rates are given to whole documents, but we may be interested in attributing the sentiment to individual sentences [Kotzias et al., 2015].”
Reducing Non-Normative Text Generation from Language Models (2020, methodology): “For example, Amazon Alexa, a virtual assistant AI, was reported to suggest a user commit suicide.”
Toward a Quantitative Survey of Dimension Reduction Techniques (2019, neutral): “Bank Marketing (bank) [33]: Direct marketing campaign data of a Portuguese bank used to predict whether a client will subscribe to a banking product or not; CIFAR10 (cifar10) [34]: Standard Computer Vision research dataset consisting of images of animals and vehicles, used for training image classfi”