Tumor segmentation from multi-modal brain MRI images is a challenging task due to the limited samples, high variance in shapes and uneven distribution of tumor morphology. The performance of automated medical image segmentation has been significant improvement by the recent advances in deep learning. However, the model predictions have not yet reached the desired level for clinical use in terms of accuracy and generalizability. In order to address the distinct problems presented in Challenges 1, 2, and 3 of BraTS 2023, we have constructed an optimization framework based on a 3D U-Net model for brain tumor segmentation. This framework incorporates a range of techniques, including various pre-processing and post-processing techniques, and transfer learning. On the validation datasets, this multi-modality brain tumor segmentation framework achieves an average lesion-wise Dice score of 0.79, 0.72, 0.74 on Challenges 1, 2, 3 respectively.
The availability of data is limited in some fields, especially for object detection tasks, where it is necessary to have correctly labeled bounding boxes around each object. A notable example of such data scarcity is found in the domain of marine biology, where it is useful to develop methods to automatically detect submarine species for environmental monitoring. To address this data limitation, the state-of-the-art machine learning strategies employ two main approaches. The first involves pretraining models on existing datasets before generalizing to the specific domain of interest. The second strategy is to create synthetic datasets specifically tailored to the target domain using methods like copy-paste techniques or ad-hoc simulators. The first strategy often faces a significant domain shift, while the second demands custom solutions crafted for the specific task. In response to these challenges, here we propose a transfer learning framework that is valid for a generic scenario. In this framework, generated images help to improve the performances of an object detector in a few-real data regime. This is achieved through a diffusion-based generative model that was pretrained on large generic datasets. With respect to the state-of-the-art, we find that it is not necessary to fine tune the generative model on the specific domain of interest. We believe that this is an important advance because it mitigates the labor-intensive task of manual labeling the images in object detection tasks. We validate our approach focusing on fishes in an underwater environment, and on the more common domain of cars in an urban setting. Our method achieves detection performance comparable to models trained on thousands of images, using only a few hundreds of input data. Our results pave the way for new generative AI-based protocols for machine learning applications in various domains.
Low-light conditions and occluded scenarios impede object detection in real-world Internet of Things (IoT) applications like autonomous vehicles and security systems. While advanced machine learning models strive for accuracy, their computational demands clash with the limitations of resource-constrained devices, hampering real-time performance. In our current research, we tackle this challenge, by introducing ``YOLO Phantom", one of the smallest YOLO models ever conceived. YOLO Phantom utilizes the novel Phantom Convolution block, achieving comparable accuracy to the latest YOLOv8n model while simultaneously reducing both parameters and model size by 43\%, resulting in a significant 19\% reduction in Giga Floating-Point Operations (GFLOPs). YOLO Phantom leverages transfer learning on our multimodal RGB-infrared dataset to address low-light and occlusion issues, equipping it with robust vision under adverse conditions. Its real-world efficacy is demonstrated on an IoT platform with advanced low-light and RGB cameras, seamlessly connecting to an AWS-based notification endpoint for efficient real-time object detection. Benchmarks reveal a substantial boost of 17\% and 14\% in frames per second (FPS) for thermal and RGB detection, respectively, compared to the baseline YOLOv8n model. For community contribution, both the code and the multimodal dataset are available on GitHub.
Due to the complex nature of human emotions and the diversity of emotion representation methods in humans, emotion recognition is a challenging field. In this research, three input modalities, namely text, audio (speech), and video, are employed to generate multimodal feature vectors. For generating features for each of these modalities, pre-trained Transformer models with fine-tuning are utilized. In each modality, a Transformer model is used with transfer learning to extract feature and emotional structure. These features are then fused together, and emotion recognition is performed using a classifier. To select an appropriate fusion method and classifier, various feature-level and decision-level fusion techniques have been experimented with, and ultimately, the best model, which combines feature-level fusion by concatenating feature vectors and classification using a Support Vector Machine on the IEMOCAP multimodal dataset, achieves an accuracy of 75.42%. Keywords: Multimodal Emotion Recognition, IEMOCAP, Self-Supervised Learning, Transfer Learning, Transformer.
Knowledge distillation is a simple but powerful way to transfer knowledge between a teacher model to a student model. Existing work suffers from at least one of the following key limitations in terms of direction and scope of transfer which restrict its use: all knowledge is transferred from teacher to student regardless of whether or not that knowledge is useful, the student is the only one learning in this exchange, and typically distillation transfers knowledge only from a single teacher to a single student. We formulate a novel form of knowledge distillation in which many models can act as both students and teachers which we call cooperative distillation. The models cooperate as follows: a model (the student) identifies specific deficiencies in it's performance and searches for another model (the teacher) who encodes learned knowledge into instructional virtual instances via counterfactual instance generation. Because different models may have different strengths and weaknesses, all models can act as either students or teachers (cooperation) when appropriate and only distill knowledge in areas specific to their strengths (focus). Since counterfactuals as a paradigm are not tied to any specific algorithm, we can use this method to distill knowledge between learners of different architectures, algorithms, and even feature spaces. We demonstrate that our approach not only outperforms baselines such as transfer learning, self-supervised learning, and multiple knowledge distillation algorithms on several datasets, but it can also be used in settings where the aforementioned techniques cannot.
Point clouds and high-resolution 3D data have become increasingly important in various fields, including surveying, construction, and virtual reality. However, simply having this data is not enough; to extract useful information, semantic labeling is crucial. In this context, we propose a method to transfer annotations from a labeled to an unlabeled point cloud using an octree structure. The structure also analyses changes between the point clouds. Our experiments confirm that our method effectively transfers annotations while addressing changes. The primary contribution of this project is the development of the method for automatic label transfer between two different point clouds that represent the same real-world object. The proposed method can be of great importance for data-driven deep learning algorithms as it can also allow circumventing stochastic transfer learning by deterministic label transfer between datasets depicting the same objects.
Background: Chest X-ray imaging-based abnormality localization, essential in diagnosing various diseases, faces significant clinical challenges due to complex interpretations and the growing workload of radiologists. While recent advances in deep learning offer promising solutions, there is still a critical issue of domain inconsistency in cross-domain transfer learning, which hampers the efficiency and accuracy of diagnostic processes. This study aims to address the domain inconsistency problem and improve autonomic abnormality localization performance of heterogeneous chest X-ray image analysis, by developing a self-supervised learning strategy called "BarlwoTwins-CXR". Methods: We utilized two publicly available datasets: the NIH Chest X-ray Dataset and the VinDr-CXR. The BarlowTwins-CXR approach was conducted in a two-stage training process. Initially, self-supervised pre-training was performed using an adjusted Barlow Twins algorithm on the NIH dataset with a Resnet50 backbone pre-trained on ImageNet. This was followed by supervised fine-tuning on the VinDr-CXR dataset using Faster R-CNN with Feature Pyramid Network (FPN). Results: Our experiments showed a significant improvement in model performance with BarlowTwins-CXR. The approach achieved a 3% increase in mAP50 accuracy compared to traditional ImageNet pre-trained models. In addition, the Ablation CAM method revealed enhanced precision in localizing chest abnormalities. Conclusion: BarlowTwins-CXR significantly enhances the efficiency and accuracy of chest X-ray image-based abnormality localization, outperforming traditional transfer learning methods and effectively overcoming domain inconsistency in cross-domain scenarios. Our experiment results demonstrate the potential of using self-supervised learning to improve the generalizability of models in medical settings with limited amounts of heterogeneous data.
Large pre-trained language models have recently been expanded and applied to programming language tasks with great success, often through further pre-training of a strictly-natural language model--where training sequences typically contain both natural and (linearised) programming language. Such approaches effectively map both modalities of the sequence into the same embedding space. However, programming language keywords (e.g. "while") often have very strictly defined semantics. As such, transfer learning from their natural language usage may not necessarily be beneficial to their code application and vise versa. Assuming an already pre-trained language model, in this work we investigate how sequence tokens can be adapted and represented differently, depending on which modality they belong to, and to the ultimate benefit of the downstream task. We experiment with separating embedding spaces between modalities during further model pre-training with modality-relative training objectives. We focus on text-to-code generation and observe consistent improvements across two backbone models and two test sets, measuring pass@$k$ and a novel incremental variation.
Textbook question answering (TQA) is a challenging task in artificial intelligence due to the complex nature of context needed to answer complex questions. Although previous research has improved the task, there are still some limitations in textual TQA, including weak reasoning and inability to capture contextual information in the lengthy context. We propose a framework (PLRTQA) that incorporates the retrieval augmented generation (RAG) technique to handle the out-of-domain scenario where concepts are spread across different lessons, and utilize transfer learning to handle the long context and enhance reasoning abilities. Our architecture outperforms the baseline, achieving an accuracy improvement of 4. 12% in the validation set and 9. 84% in the test set for textual multiple-choice questions. While this paper focuses on solving challenges in the textual TQA, It provides a foundation for future work in multimodal TQA where the visual components are integrated to address more complex educational scenarios. Code: https://github.com/hessaAlawwad/PLR-TQA
Graph transformers typically lack third-order interactions, limiting their geometric understanding which is crucial for tasks like molecular geometry prediction. We propose the Triplet Graph Transformer (TGT) that enables direct communication between pairs within a 3-tuple of nodes via novel triplet attention and aggregation mechanisms. TGT is applied to molecular property prediction by first predicting interatomic distances from 2D graphs and then using these distances for downstream tasks. A novel three-stage training procedure and stochastic inference further improve training efficiency and model performance. Our model achieves new state-of-the-art (SOTA) results on open challenge benchmarks PCQM4Mv2 and OC20 IS2RE. We also obtain SOTA results on QM9, MOLPCBA, and LIT-PCBA molecular property prediction benchmarks via transfer learning. We also demonstrate the generality of TGT with SOTA results on the traveling salesman problem (TSP).
Dataset distillation (DD) has emerged as a widely adopted technique for crafting a synthetic dataset that captures the essential information of a training dataset, facilitating the training of accurate neural models. Its applications span various domains, including transfer learning, federated learning, and neural architecture search. The most popular methods for constructing the synthetic data rely on matching the convergence properties of training the model with the synthetic dataset and the training dataset. However, using the empirical loss as the criterion must be thought of as auxiliary in the same sense that the training set is an approximate substitute for the population distribution, and the latter is the data of interest. Yet despite its popularity, an aspect that remains unexplored is the relationship of DD to its generalization, particularly across uncommon subgroups. That is, how can we ensure that a model trained on the synthetic dataset performs well when faced with samples from regions with low population density? Here, the representativeness and coverage of the dataset become salient over the guaranteed training error at inference. Drawing inspiration from distributionally robust optimization, we introduce an algorithm that combines clustering with the minimization of a risk measure on the loss to conduct DD. We provide a theoretical rationale for our approach and demonstrate its effective generalization and robustness across subgroups through numerical experiments. The source code is available at https://github.com/Mming11/RobustDatasetDistillation.
Potato plants are plants that are beneficial to humans. Like other plants in general, potato plants also have diseases; if this disease is not treated immediately, there will be a significant decrease in food production. Therefore, it is necessary to detect diseases quickly and precisely so that disease control can be carried out effectively and efficiently. Classification of potato leaf disease can be done directly. Still, the symptoms cannot always explain the type of disease that attacks potato leaves because there are many types of diseases with symptoms that look the same. Humans also have deficiencies in determining the results of identification of potato leaf disease, so sometimes the results of identification between individuals can be different. Therefore, the use of Deep Learning for the classification process of potato leaf disease is expected to shorten the time and have a high classification accuracy. This study uses a deep learning method with the DenseNet201 architecture. The choice to use the DenseNet201 algorithm in this study is because the model can identify important features of potato leaves and recognize early signs of emerging diseases. This study aimed to evaluate the effectiveness of the transfer learning method with the DenseNet201 architecture in increasing the classification accuracy of potato leaf disease compared to traditional classification methods. This study uses two types of scenarios, namely, comparing the number of dropouts and comparing the three optimizers. This test produces the best model using dropout 0.1 and Adam optimizer with an accuracy of 99.5% for training, 95.2% for validation, and 96% for the confusion matrix. In this study, using data testing, as many as 40 images were tested into the model that has been built. The test results on this model resulted in a new accuracy for classifying potato leaf disease, namely 92.5%.
Scaling laws provide important insights that can guide the design of large language models (LLMs). Existing work has primarily focused on studying scaling laws for pretraining (upstream) loss. However, in transfer learning settings, in which LLMs are pretrained on an unsupervised dataset and then finetuned on a downstream task, we often also care about the downstream performance. In this work, we study the scaling behavior in a transfer learning setting, where LLMs are finetuned for machine translation tasks. Specifically, we investigate how the choice of the pretraining data and its size affect downstream performance (translation quality) as judged by two metrics: downstream cross-entropy and BLEU score. Our experiments indicate that the size of the finetuning dataset and the distribution alignment between the pretraining and downstream data significantly influence the scaling behavior. With sufficient alignment, both downstream cross-entropy and BLEU score improve monotonically with more pretraining data. In such cases, we show that it is possible to predict the downstream BLEU score with good accuracy using a log-law. However, there are also cases where moderate misalignment causes the BLEU score to fluctuate or get worse with more pretraining, whereas downstream cross-entropy monotonically improves. By analyzing these observations, we provide new practical insights for choosing appropriate pretraining data.
Glioblastoma is a highly malignant brain tumor with a life expectancy of only 3 to 6 months without treatment. Detecting and predicting its survival and grade accurately are crucial. This study introduces a novel approach using transfer learning techniques. Various pre-trained networks, including EfficientNet, ResNet, VGG16, and Inception, were tested through exhaustive optimization to identify the most suitable architecture. Transfer learning was applied to fine-tune these models on a glioblastoma image dataset, aiming to achieve two objectives: survival and tumor grade prediction.The experimental results show 65% accuracy in survival prediction, classifying patients into short, medium, or long survival categories. Additionally, the prediction of tumor grade achieved an accuracy of 97%, accurately differentiating low-grade gliomas (LGG) and high-grade gliomas (HGG). The success of the approach is attributed to the effectiveness of transfer learning, surpassing the current state-of-the-art methods. In conclusion, this study presents a promising method for predicting the survival and grade of glioblastoma. Transfer learning demonstrates its potential in enhancing prediction models, particularly in scenarios with limited large datasets. These findings hold promise for improving diagnostic and treatment approaches for glioblastoma patients.
Recent advancements in large language models have sparked interest in their extraordinary and near-superhuman capabilities, leading researchers to explore methods for evaluating and optimizing these abilities, which is called superalignment. In this context, our paper delves into the realm of vision foundation models, focusing on the concept of weak-to-strong generalization, which involves using a weaker model to supervise a stronger one, aiming to enhance the latter's capabilities beyond the former's limits. We introduce a novel and adaptively adjustable loss function for weak-to-strong supervision. Our comprehensive experiments span various scenarios, including few-shot learning, transfer learning, noisy label learning, and common knowledge distillation settings. The results are striking: our approach not only exceeds the performance benchmarks set by strong-to-strong generalization but also surpasses the outcomes of fine-tuning strong models with whole datasets. This compelling evidence underscores the significant potential of weak-to-strong generalization, showcasing its capability to substantially elevate the performance of vision foundation models. The code is available at https://github.com/ggjy/vision_weak_to_strong.
To handle AI tasks that combine perception and logical reasoning, recent work introduces Neurosymbolic Deep Neural Networks (NS-DNNs), which contain -- in addition to traditional neural layers -- symbolic layers: symbolic expressions (e.g., SAT formulas, logic programs) that are evaluated by symbolic solvers during inference. We identify and formalize an intuitive, high-level principle that can guide the design and analysis of NS-DNNs: symbol correctness, the correctness of the intermediate symbols predicted by the neural layers with respect to a (generally unknown) ground-truth symbolic representation of the input data. We demonstrate that symbol correctness is a necessary property for NS-DNN explainability and transfer learning (despite being in general impossible to train for). Moreover, we show that the framework of symbol correctness provides a precise way to reason and communicate about model behavior at neural-symbolic boundaries, and gives insight into the fundamental tradeoffs faced by NS-DNN training algorithms. In doing so, we both identify significant points of ambiguity in prior work, and provide a framework to support further NS-DNN developments.
This paper presents an effective transfer learning framework for language adaptation in text-to-speech systems, with a focus on achieving language adaptation using minimal labeled and unlabeled data. While many works focus on reducing the usage of labeled data, very few consider minimizing the usage of unlabeled data. By utilizing self-supervised features in the pretraining stage, replacing the noisy portion of pseudo labels with these features during fine-tuning, and incorporating an embedding initialization trick, our method leverages more information from unlabeled data compared to conventional approaches. Experimental results show that our framework is able to synthesize intelligible speech in unseen languages with only 4 utterances of labeled data and 15 minutes of unlabeled data. Our methodology continues to surpass conventional techniques, even when a greater volume of data is accessible. These findings highlight the potential of our data-efficient language adaptation framework.
Accurately modeling the protein fitness landscapes holds great importance for protein engineering. Recently, due to their capacity and representation ability, pre-trained protein language models have achieved state-of-the-art performance in predicting protein fitness without experimental data. However, their predictions are limited in accuracy as well as interpretability. Furthermore, such deep learning models require abundant labeled training examples for performance improvements, posing a practical barrier. In this work, we introduce FSFP, a training strategy that can effectively optimize protein language models under extreme data scarcity. By combining the techniques of meta-transfer learning, learning to rank, and parameter-efficient fine-tuning, FSFP can significantly boost the performance of various protein language models using merely tens of labeled single-site mutants from the target protein. The experiments across 87 deep mutational scanning datasets underscore its superiority over both unsupervised and supervised approaches, revealing its potential in facilitating AI-guided protein design.
Lung cancer (LC) remains a leading cause of death worldwide. Early diagnosis is critical to protect innocent human lives. Computed tomography (CT) scans are one of the primary imaging modalities for lung cancer diagnosis. However, manual CT scan analysis is time-consuming and prone to errors/not accurate. Considering these shortcomings, computational methods especially machine learning and deep learning algorithms are leveraged as an alternative to accelerate the accurate detection of CT scans as cancerous, and non-cancerous. In the present article, we proposed a novel transfer learning-based predictor called, Lung-EffNet for lung cancer classification. Lung-EffNet is built based on the architecture of EfficientNet and further modified by adding top layers in the classification head of the model. Lung-EffNet is evaluated by utilizing five variants of EfficientNet i. e., B0–B4. The experiments are conducted on the benchmark dataset “IQ-OTH/NCCD” for lung cancer patients grouped as benign, malignant, or normal based on the presence or absence of lung cancer. The class imbalance issue was handled through multiple data augmentation methods to overcome the biases. The developed model Lung-EffNet attained 99.10% of accuracy and a score of 0.97 to 0.99 of ROC on the test set. We compared the efficacy of the proposed fine-tuned pre-trained EfficientNet with other pre-trained CNN architectures. The predicted outcomes demonstrate that EfficientNetB1 based Lung-EffNet outperforms other CNNs in terms of both accuracy and efficiency. Moreover, it is faster and requires fewer parameters to train than other CNN based models, making it a good choice for large-scale deployment in clinical settings and a promising tool for automated lung cancer diagnosis from CT scan images.
Mainstream cross-lingual task-oriented dialogue (ToD) systems leverage the transfer learning paradigm by training a joint model for intent recognition and slot-filling in English and applying it, zero-shot, to other languages. We address a gap in prior research, which often overlooked the transfer to lower-resource colloquial varieties due to limited test data. Inspired by prior work on English varieties, we craft and manually evaluate perturbation rules that transform German sentences into colloquial forms and use them to synthesize test sets in four ToD datasets. Our perturbation rules cover 18 distinct language phenomena, enabling us to explore the impact of each perturbation on slot and intent performance. Using these new datasets, we conduct an experimental evaluation across six different transformers. Here, we demonstrate that when applied to colloquial varieties, ToD systems maintain their intent recognition performance, losing 6% (4.62 percentage points) in accuracy on average. However, they exhibit a significant drop in slot detection, with a decrease of 31% (21 percentage points) in slot F1 score. Our findings are further supported by a transfer experiment from Standard American English to synthetic Urban African American Vernacular English.