TIML: Task-Informed Meta-Learning for Agriculture

Tseng, Gabriel; Kerner, Hannah; Rolnick, David

doi:10.48550/arxiv.2202.02124

Cited by 3 publications

(5 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Obtaining MS time series of large territories can be prohibitive due to the cost and time required to acquire them [85], [86]. In addition, no previous work has explored the possibility of adding ancillary data to enhance the spectral unmixing results, which are used successfully in other computer vision tasks [11], [12], [13].…”

Section: Discussionmentioning

confidence: 99%

“…In parallel, there exist few works that incorporate ancillary data to improve the performance of DL models. Most of these studies occur in the field of computer vision, such as high inter-class similarity classification problems [11], plankton image classification [12] or crop type mapping [13].…”

Section: B Spectral Unmixingmentioning

confidence: 99%

“…In contrast to traditional methods, the application of DL in SU facilitates the exploitation of ancillary information such as geographic location, topography, and climate. For example, in the field of computer vision, ancillary data has been successfully used by DL models to improve the performance during image classification [11], [12], [13]. However, the introduction of ancillary information remains unexplored in spectral unmixing methods.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Bidirectional Recurrent Imputation and Abundance Estimation of LULC Classes With MODIS Multispectral Time-Series and Geo-Topographic and Climatic Data

Rodríguez-Ortega,

Khaldi,

Alcaraz-Segura

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Remotely sensed data are dominated by mixed Land Use and Land Cover (LULC) types. Spectral unmixing (SU) is a key technique that disentangles mixed pixels into constituent LULC types and their abundance fractions. While existing studies on Deep Learning (DL) for SU typically focus on single time-step hyperspectral (HS) or multispectral (MS) data, our work pioneers SU using MODIS MS time series, addressing missing data with end-to-end DL models. Our approach enhances a Long-Short Term Memory (LSTM)-based model by incorporating geographic, topographic (geo-topographic), and climatic ancillary information. Notably, our method eliminates the need for explicit endmember extraction, instead learning the input-output relationship between mixed spectra and LULC abundances through supervised learning. Experimental results demonstrate that integrating spectral-temporal input data with geo-topographic and climatic information significantly improves the estimation of LULC abundances in mixed pixels. To facilitate this study, we curated a novel labeled dataset for Andalusia (Spain) with monthly MODIS multispectral time series at 460m resolution for 2013. Named Andalusia MultiSpectral MultiTemporal Unmixing (Andalusia-MSMTU), this dataset provides pixellevel annotations of LULC abundances along with ancillary information. The dataset 1 and code 2 are available to the public.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: B Spectral Unmixingmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Bidirectional Recurrent Imputation and Abundance Estimation of LULC Classes With MODIS Multispectral Time-Series and Geo-Topographic and Climatic Data

Rodríguez-Ortega,

Khaldi,

Alcaraz-Segura

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Few-Shot Learning Metric-based (Li and Chao, 2020;Arg¨ueso et al, 2020;Monowar et al, 2022;Egusquiza et al, 2022;Li and Yang, 2020;Zhang et al, 2022a;Cai et al, 2021) Model-based ) Optimization-based Tseng et al, 2022;Zhai et al, 2022;Saleem et al, 2020…”

Section: Methodsmentioning

confidence: 99%

Meta‐learning shows great potential in plant disease recognition under few available samples

Deng

Wang

et al. 2023

The Plant Journal

View full text Add to dashboard Cite

Plant diseases worsen the threat of food shortage with the growing global population, and disease recognition is the basis for the effective prevention and control of plant diseases. Deep learning has made significant breakthroughs in the field of plant disease recognition. Compared with traditional deep learning, metalearning can still maintain more than 90% accuracy in disease recognition with small samples. However, there is no comprehensive review on the application of meta-learning in plant disease recognition. Here, we mainly summarize the functions, advantages, and limitations of meta-learning research methods and their applications for plant disease recognition with a few data scenarios. Finally, we outline several research avenues for utilizing current and future meta-learning in plant science. This review may help plant science researchers obtain faster, more accurate, and more credible solutions through deep learning with fewer labeled samples.

show abstract

“…It focuses on using previous knowledge to increase learning efficiency and generalization. More interestingly, the authors of [67] proposed TIML: Task-Informed Meta-Learning for Agriculture. The primary goal of this study is to investigate how model-agnostic meta-learning (MAML) weights can be modulated, even when all tasks were selected from a single data set.…”

Section: Highlymentioning

confidence: 99%

Remote Sensing for Agriculture in the Era of Industry 5.0—A Survey

Victor,

Maddikunta,

Mary

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Agriculture can be regarded as the backbone of human civilization. As technology evolved, the synergy between agriculture and remote sensing has brought about a paradigm shift, thereby entirely revolutionizing the traditional agricultural practices. Nevertheless, the adoption of remote sensing technologies in agriculture face various challenges in terms of limited spatial and temporal coverage, high cloud cover, low data quality and so on. Industry 5.0 marks a new era in the industrial revolution, where humans and machines collaborate closely, leveraging their distinct capabilities, thereby enhancing the decision making capabilities, sustainability and resilience. This paper provides a comprehensive survey on remote sensing technologies and related aspects in dealing with the various agricultural practices in the Industry 5.0 (I5.0) era. We also elaborately discuss the various applications pertaining to I5.0enabled remote sensing for agriculture. Finally, we discuss several challenges and issues related to the integration of I5.0 technologies in agricultural remote sensing. This comprehensive survey on remote sensing for agriculture in Industry 5.0 era offers valuable insights into the current state, challenges, and potential advancements in the integration of remote sensing technologies and Industry 5.0 principles in agriculture, thus paving the way for future research, development, and implementation strategies in this domain.

show abstract

TIML: Task-Informed Meta-Learning for Agriculture

Cited by 3 publications

References 18 publications

Bidirectional Recurrent Imputation and Abundance Estimation of LULC Classes With MODIS Multispectral Time-Series and Geo-Topographic and Climatic Data

Bidirectional Recurrent Imputation and Abundance Estimation of LULC Classes With MODIS Multispectral Time-Series and Geo-Topographic and Climatic Data

Meta‐learning shows great potential in plant disease recognition under few available samples

Remote Sensing for Agriculture in the Era of Industry 5.0—A Survey

Contact Info

Product

Resources

About