Filters








22,264 Hits in 6.9 sec

Shot in the Dark: Few-Shot Learning with No Base-Class Labels [article]

Zitian Chen, Subhransu Maji, Erik Learned-Miller
2021 arXiv   pre-print
In this work, we show that, surprisingly, off-the-shelf self-supervised learning outperforms transductive few-shot methods by 3.9% for 5-shot accuracy on miniImageNet without using any base class labels  ...  Few-shot learning aims to build classifiers for new classes from a small number of labeled examples and is commonly facilitated by access to examples from a distinct set of 'base classes'.  ...  The UBC learning shows competitive performance against labeled-base-class learning in the few-shot setting.  ... 
arXiv:2010.02430v2 fatcat:v2i4a3pkzfdqjpik3arzfxtpmy

Closing the Generalization Gap in One-Shot Object Detection [article]

Claudio Michaelis, Matthias Bethge, Alexander S. Ecker
2020 arXiv   pre-print
Despite substantial progress in object detection and few-shot learning, detecting objects based on a single example - one-shot object detection - remains a challenge: trained models exhibit a substantial  ...  Our results suggest that the key to strong few-shot detection models may not lie in sophisticated metric learning approaches, but instead in scaling the number of categories.  ...  Few-shot learning The two most common approaches to few-shot learning have been, broadly speaking, based on metric learning [22, 48, 42] and meta learning: Learn a good way to learn a new task [12,  ... 
arXiv:2011.04267v1 fatcat:g7rr7624znagficq5or4xyzrkq

"Diversity and Uncertainty in Moderation" are the Key to Data Selection for Multilingual Few-shot Transfer [article]

Shanu Kumar, Sandipan Dandapat, Monojit Choudhury
2022 arXiv   pre-print
Our experiments show that the gradient and loss embedding-based strategies consistently outperform random data selection baselines, with gains varying with the initial performance of the zero-shot transfer  ...  This paper explores various strategies for selecting data for annotation that can result in a better few-shot transfer.  ...  As sentences with similar context will also have similar class labels in the case of POS and NER tasks, further decreasing the diversity in samples.  ... 
arXiv:2206.15010v1 fatcat:js53kwvhyjewzkcnvyhb5v5x54

Large-scale zero-shot learning in the wild: Classifying zoological illustrations

Lise Stork, Andreas Weber, Jaap van den Herik, Aske Plaat, Fons Verbeek, Katherine Wolstencroft
2021 Ecological Informatics  
We explore zero-shot learning to address the problem, where features are learned from classes with medium to large samples, which are then transferred to recognise classes with few or no training samples  ...  (BHL), can be used to share knowledge between classes for zero-shot learning.  ...  Nguyen and Sophia Ananiadou (National Centre for Text Mining) for providing the text embeddings from the Biodiversity Heritage Library.  ... 
doi:10.1016/j.ecoinf.2021.101222 fatcat:jxcutjpsvrfmdhxsr4egqnqjie

ATZSL: Defensive Zero-Shot Recognition in the Presence of Adversaries [article]

Xingxing Zhang, Shupeng Gui, Zhenfeng Zhu, Yao Zhao, Ji Liu
2019 arXiv   pre-print
Additionally, our framework can be extended to deal with the poisoned scenario of unseen class attributes.  ...  Zero-shot learning (ZSL) has received extensive attention recently especially in areas of fine-grained object recognition, retrieval, and image captioning.  ...  The goal is to recognize data belonging to the classes that have no labeled samples.  ... 
arXiv:1910.10994v2 fatcat:ydc7fgcawve2xnqiqpaidmaa7e

Toward zero-shot Entity Recognition in Task-oriented Conversational Agents

Marco Guerini, Simone Magnolini, Vevake Balaraman, Bernardo Magnini
2018 Proceedings of the 19th Annual SIGdial Meeting on Discourse and Dialogue  
We present a domain portable zero-shot learning approach for entity recognition in task-oriented conversational agents, which does not assume any annotated sentences at training time.  ...  Rather, we derive a neural model of the entity names based only on available gazetteers, and then apply the model to recognize new entities in the context of user utterances.  ...  The authors thank the anonymous reviewers and Hendrik Buschmeier for their help and suggestions.  ... 
doi:10.18653/v1/w18-5036 dblp:conf/sigdial/GueriniMBM18 fatcat:cvv6j7iqzreetccbkdirhmcyly

Going Deeper into Recognizing Actions in Dark Environments: A Comprehensive Benchmark Study [article]

Yuecong Xu, Jianfei Yang, Haozhi Cao, Jianxiong Yin, Zhenghua Chen, Xiaoli Li, Zhengguo Li, Qianwen Xu
2022 arXiv   pre-print
to tackle the challenge of AR in dark environments.  ...  To dive deeper into exploring solutions for AR in dark environments, we launched the UG2+ Challenge Track 2 (UG2-2) in IEEE CVPR 2021, with a goal of evaluating and advancing the robustness of AR models  ...  One of which is few-shot learning (Bo, Lu, & He, 2020; Kumar Dwivedi, Gupta, Mitra, Ahmed, & Jain, 2019) , which has enabled models to be trained with limited data while generalizing to unseen test data  ... 
arXiv:2202.09545v2 fatcat:4bw7ntifg5faxbbiaccuemublm

Multimodal Disentanglement Variational AutoEncoders for Zero-Shot Cross-Modal Retrieval

Jialin Tian, Kai Wang, Xing Xu, Zuo Cao, Fumin Shen, Heng Tao Shen
2022 Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval  
with seen classes in the training set.  ...  Zero-Shot Cross-Modal Retrieval (ZS-CMR) has recently drawn increasing attention as it focuses on a practical retrieval scenario, i.e., the multimodal test set consists of unseen classes that are disjoint  ...  Zero-Shot Sketch-Based Image Retrieval.  ... 
doi:10.1145/3477495.3532028 fatcat:axwwf2jxufbw5n75n4kooacery

Single-Shot Convolution Neural Networks for Real-Time Fruit Detection Within the Tree

Kushtrim Bresilla, Giulio Demetrio Perulli, Alexandra Boini, Brunella Morandi, Luca Corelli Grappadelli, Luigi Manfrini
2019 Frontiers in Plant Science  
We want this confidence score to be high if a fruit exists in a cell, otherwise to be zero, if no fruit is in the cell. More than 100 images of apple and pear trees were taken.  ...  Labeling images for training consisted on manually specifying the bounding boxes for fruits, where (x, y) are the center coordinates of the box and (w, h) are width and height.  ...  In a dataset with C class labels, the output tensor is S × S × (C + B × 5).  ... 
doi:10.3389/fpls.2019.00611 pmid:31178875 pmcid:PMC6537632 fatcat:vopqc3lrhrb5vgl43gllbdve7y

Image Classification in the Dark using Quanta Image Sensors [article]

Abhiram Gnanasambandam, Stanley H. Chan
2020 arXiv   pre-print
However, in dark environments when the photon flux is low, image classification becomes difficult because the measured signal is suppressed by noise.  ...  We show that with student-teacher learning, we are able to achieve image classification at a photon level of one photon per pixel or lower.  ...  This, in turn, forces the network to "denoise" the shot noise and read noise in x QIS before predicting the label.  ... 
arXiv:2006.02026v3 fatcat:dqeg2h4lzra2bjvqz3ywqu5p4e

Mass classification of dark matter perturbers of stellar tidal streams [article]

Francesco Montanari, Juan García-Bellido
2021 arXiv   pre-print
While the schemes do not assume a specific dark matter model, we are mainly interested in discerning the primordial black holes cold dark matter (PBH CDM) hypothesis form the standard particle dark matter  ...  Stellar streams formed by tidal stripping of progenitors orbiting around the Milky Way are expected to be perturbed by encounters with dark matter subhalos.  ...  Data Availability The code underlying this article is available at https://gitlab.com/montanari/ stream-dm.  ... 
arXiv:2012.11482v2 fatcat:dmudt626pjc2bpaz676d64bd2a

ARID: A Comprehensive Study on Recognizing Actions in the Dark and A New Benchmark Dataset [article]

Yuecong Xu, Jianfei Yang, Haozhi Cao, Kezhi Mao, Jianxiong Yin, Simon See
2021 arXiv   pre-print
Though progress has been made in the action recognition task for videos in normal illumination, few have studied action recognition in the dark.  ...  We bridge the gap of the lack of data for this task by collecting a new dataset: the Action Recognition in the Dark (ARID) dataset. It consists of over 3,780 video clips with 11 action categories.  ...  The lighting condition of each scene is different, with no direct light shot on the actor in almost all videos.  ... 
arXiv:2006.03876v3 fatcat:ty4qtfp5ujddzdxqw3t4qhxiwi

Seeing into Darkness: Scotopic Visual Recognition

Bo Chen, Pietro Perona
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Shot noise: The number of photons incident on a pixel i in the unit time follows a Poisson distribution whose rate (in Hz) depends on both the pixel intensity I i ∈ [0, 1] and a dark current ǫ dc : P (  ...  As photon-counting sensors are designed to have low read noise and low fixed pattern noise [12, 46, 13] , we focus on modeling the shot noise and dark current only.  ... 
doi:10.1109/cvpr.2017.771 dblp:conf/cvpr/ChenP17 fatcat:qc4dgs3euva2jgpyhookrjzhfm

Seeing into Darkness: Scotopic Visual Recognition [article]

Bo Chen, Pietro Perona
2016 arXiv   pre-print
When photons are few and far in between, the concept of 'image' breaks down and it is best to consider directly the flow of photons.  ...  Computer vision in this regime, which we call 'scotopic', is radically different from the classical image-based paradigm in that visual computations (classification, control, search) have to take place  ...  Assume that a probabilistic model is available to predict the class label C given a sensory input X 1:t of any duration t -either provided by the application or learned from labeled data using techniques  ... 
arXiv:1610.00405v1 fatcat:lfyioirfnngwnm3k2mjh4sw47e

Illuminating the dark spaces of healthcare with ambient intelligence

Albert Haque, Arnold Milstein, Li Fei-Fei
2020 Nature  
Advances in machine learning and contactless sensors have given rise to ambient intelligence-physical spaces that are sensitive and responsive to the presence of humans.  ...  In daily living spaces, ambient intelligence could prolong the independence of older individuals and improve the management of individuals with a chronic disease by understanding everyday behaviour.  ...  For low-resource healthcare providers, few-shot learning-algorithms capable of learning from as few as one or two examples 144 -could be used.  ... 
doi:10.1038/s41586-020-2669-y pmid:32908264 fatcat:fj5mnqkk4nglnfh3scmm7y6zou
« Previous Showing results 1 — 15 out of 22,264 results