More Like this

1. A Multi-Level Optimization Framework for End-to-End Text Augmentation

   https://doi.org/10.1162/tacl_a_00464  

   Somayajula, Sai Ashish; Song, Linfeng; Xie, Pengtao (January 2022, Transactions of the Association for Computational Linguistics)

   Abstract Text augmentation is an effective technique in alleviating overfitting in NLP tasks. In existing methods, text augmentation and downstream tasks are mostly performed separately. As a result, the augmented texts may not be optimal to train the downstream model. To address this problem, we propose a three-level optimization framework to perform text augmentation and the downstream task end-to- end. The augmentation model is trained in a way tailored to the downstream task. Our framework consists of three learning stages. A text summarization model is trained to perform data augmentation at the first stage. Each summarization example is associated with a weight to account for its domain difference with the text classification data. At the second stage, we use the model trained at the first stage to perform text augmentation and train a text classification model on the augmented texts. At the third stage, we evaluate the text classification model trained at the second stage and update weights of summarization examples by minimizing the validation loss. These three stages are performed end-to-end. We evaluate our method on several text classification datasets where the results demonstrate the effectiveness of our method. Code is available at https://github.com/Sai-Ashish/End-to-End-Text-Augmentation. 

   more » « less
2. Data science competition for cross-site individual tree species identification from airborne remote sensing data

   https://doi.org/10.7717/peerj.16578  

   Graves, Sarah J; Marconi, Sergio; Stewart, Dylan; Harmon, Ira; Weinstein, Ben; Kanazawa, Yuzi; Scholl, Victoria M; Joseph, Maxwell B; McGlinchy, Joseph; Browne, Luke; et al (January 2023, PeerJ)

   Data on individual tree crowns from remote sensing have the potential to advance forest ecology by providing information about forest composition and structure with a continuous spatial coverage over large spatial extents. Classifying individual trees to their taxonomic species over large regions from remote sensing data is challenging. Methods to classify individual species are often accurate for common species, but perform poorly for less common species and when applied to new sites. We ran a data science competition to help identify effective methods for the task of classification of individual crowns to species identity. The competition included data from three sites to assess each methods’ ability to generalize patterns across two sites simultaneously and apply methods to an untrained site. Three different metrics were used to assess and compare model performance. Six teams participated, representing four countries and nine individuals. The highest performing method from a previous competition in 2017 was applied and used as a baseline to understand advancements and changes in successful methods. The best species classification method was based on a two-stage fully connected neural network that significantly outperformed the baseline random forest and gradient boosting ensemble methods. All methods generalized well by showing relatively strong performance on the trained sites (accuracy = 0.46–0.55, macro F1 = 0.09–0.32, cross entropy loss = 2.4–9.2), but generally failed to transfer effectively to the untrained site (accuracy = 0.07–0.32, macro F1 = 0.02–0.18, cross entropy loss = 2.8–16.3). Classification performance was influenced by the number of samples with species labels available for training, with most methods predicting common species at the training sites well (maximum F1 score of 0.86) relative to the uncommon species where none were predicted. Classification errors were most common between species in the same genus and different species that occur in the same habitat. Most methods performed better than the baseline in detecting if a species was not in the training data by predicting an untrained mixed-species class, especially in the untrained site. This work has highlighted that data science competitions can encourage advancement of methods, particularly by bringing in new people from outside the focal discipline, and by providing an open dataset and evaluation criteria from which participants can learn. 

   more » « less
3. Modeling Label Space Interactions in Multi-label Classification using Box Embeddings

   Patel, Dhruvesh; Dangati, Pavitra; Lee, Jay-Yoon; Boratko, Michael; McCallum, Andrew (January 2022, ICLR 2022 Poster)

   Multi-label classification is a challenging structured prediction task in which a set of output class labels are predicted for each input. Real-world datasets often have natural or latent taxonomic relationships between labels, making it desirable for models to employ label representations capable of capturing such taxonomies. Most existing multi-label classification methods do not do so, resulting in label predictions that are inconsistent with the taxonomic constraints, thus failing to accurately represent the fundamentals of problem setting. In this work, we introduce the multi-label box model (MBM), a multi-label classification method that combines the encoding power of neural networks with the inductive bias and probabilistic semantics of box embeddings (Vilnis, et al 2018). Box embeddings can be understood as trainable Venn-diagrams based on hyper-rectangles. Representing labels by boxes rather than vectors, MBM is able to capture taxonomic relations among labels. Furthermore, since box embeddings allow these relations to be learned by stochastic gradient descent from data, and to be read as calibrated conditional probabilities, our model is endowed with a high degree of interpretability. This interpretability also facilitates the injection of partial information about label-label relationships into model training, to further improve its consistency. We provide theoretical grounding for our method and show experimentally the model's ability to learn the true latent taxonomic structure from data. Through extensive empirical evaluations on both small and large-scale multi-label classification datasets, we show that BBM can significantly improve taxonomic consistency while preserving or surpassing the state-of-the-art predictive performance. 

   more » « less
4. Self-Supervised Learning for Panoptic Segmentation of Multiple Fruit Flower Species

   https://doi.org/10.1109/LRA.2022.3217000  

   Siddique, Abubakar; Tabb, Amy; Medeiros, Henry (October 2022, IEEE Robotics and Automation Letters)

   Convolutional neural networks trained using manually generated labels are commonly used for semantic or instance segmentation. In precision agriculture, automated flower detection methods use supervised models and post-processing techniques that may not perform consistently as the appearance of the flowers and the data acquisition conditions vary. We propose a self-supervised learning strategy to enhance the sensitivity of segmentation models to different flower species using automatically generated pseudo-labels. We employ a data augmentation and refinement approach to improve the accuracy of the model predictions. The augmented semantic predictions are then converted to panoptic pseudo-labels to iteratively train a multi-task model. The self-supervised model predictions can be refined with existing post-processing approaches to further improve their accuracy. An evaluation on a multi-species fruit tree flower dataset demonstrates that our method outperforms state-of-the-art models without computationally expensive post-processing steps, providing a new baseline for flower detection applications. 

   more » « less
5. Label-Retrieval-Augmented Diffusion Models for Learning from Noisy Labels

   Chen, Jian; Zhang, Ruiyi; Yu, Tong; Sharma, Rohan; Xu, Zhiqiang; Sun, Tong; Chen, Changyou (December 2023, Curran Associates, Inc.)

   Oh, A; Naumann, T; Globerson, A; Saenko, K; Hardt, M; Levine, S (Ed.)

   Learning from noisy labels is a long-standing problem in machine learning for real applications. One of the main research lines focuses on learning a label corrector to purify potential noisy labels. However, these methods typically rely on strict assumptions and are limited to certain types of label noise. In this paper, we reformulate the label-noise problem from a generative-model perspective, i.e., labels are generated by gradually refining an initial random guess. This new perspective immediately enables existing powerful diffusion models to seamlessly learn the stochastic generative process. Once the generative uncertainty is modeled, we can perform classification inference using maximum likelihood estimation of labels. To mitigate the impact of noisy labels, we propose Label-Retrieval- Augmented (LRA) diffusion model 1, which leverages neighbor consistency to effectively construct pseudo-clean labels for diffusion training. Our model is flexible and general, allowing easy incorporation of different types of conditional information, e.g., use of pre-trained models, to further boost model performance. Extensive experiments are conducted for evaluation. Our model achieves new state-of-the-art (SOTA) results on all standard real-world benchmark datasets. Remarkably, by incorporating conditional information from the powerful CLIP model, our method can boost the current SOTA accuracy by 10-20 absolute points in many cases. 

   more » « less