Search

Segmentation of brain tumors is a critical step in treatment planning, yet manual segmentation is both time-consuming and subjective, relying heavily on the expertise of radiologists. In Sub-Saharan Africa, this challenge is magnified by overburdened medical systems and limited access to advanced imaging modalities and expert radiologists. Automating brain tumor segmentation using deep learning offers a promising solution. Convolutional Neural Networks (CNNs), especially the U-Net architecture, have shown significant potential. However, a major challenge remains: achieving generalizability across different datasets. This study addresses this gap by developing a deep learning ensemble that integrates UNet3D, V-Net, and MSA-VNet models for the semantic segmentation of gliomas. By initially training on the BraTS-GLI dataset and fine-tuning with the BraTS-SSA dataset, we enhance model performance. Our ensemble approach significantly outperforms individual models, achieving DICE scores of 0.8358 for Tumor Core, 0.8521 for Whole Tumor, and 0.8167 for Enhancing Tumor. These results underscore the potential of ensemble methods in improving the accuracy and reliability of automated brain tumor segmentation, particularly in resource-limited settings.

Although binary classification is a well-studied problem, training reliable classifiers under severe class imbalance remains a challenge. Recent techniques mitigate the ill effects of imbalance on training by modifying the loss functions or optimization methods. We observe that different hyperparameter values on these loss functions perform better at different recall values. We propose to exploit this fact by training one model over a distribution of hyperparameter values--instead of a single value--via Loss Conditional Training (LCT). Experiments show that training over a distribution of hyperparameters not only approximates the performance of several models but actually improves the overall performance of models on both CIFAR and real medical imaging applications, such as melanoma and diabetic retinopathy detection. Furthermore, training models with LCT is more efficient because some hyperparameter tuning can be conducted after training to meet individual needs without needing to retrain from scratch.

Generating commonsense assertions within a given story context remains a difficult task for modern language models. Previous research has addressed this problem by aligning commonsense inferences with stories and training language generation models accordingly. One of the challenges is determining which topic or entity in the story should be the focus of an inferred assertion. Prior approaches lack the ability to control specific aspects of the generated assertions. In this work, we introduce "hinting," a data augmentation technique that enhances contextualized commonsense inference. "Hinting" employs a prefix prompting strategy using both hard and soft prompts to guide the inference process. To demonstrate its effectiveness, we apply "hinting" to two contextual commonsense inference datasets: ParaCOMET and GLUCOSE, evaluating its impact on both general and context-specific inference. Furthermore, we evaluate "hinting" by incorporating synonyms and antonyms into the hints. Our results show that "hinting" does not compromise the performance of contextual commonsense inference while offering improved controllability., Comment: Submitted to ACL Rolling Review. arXiv admin note: text overlap with arXiv:2302.05406

Researchers in biomedical research, public health, and the life sciences often spend weeks or months discovering, accessing, curating, and integrating data from disparate sources, significantly delaying the onset of actual analysis and innovation. Instead of countless developers creating redundant and inconsistent data pipelines, BioBricks.ai offers a centralized data repository and a suite of developer-friendly tools to simplify access to scientific data. Currently, BioBricks.ai delivers over ninety biological and chemical datasets. It provides a package manager-like system for installing and managing dependencies on data sources. Each 'brick' is a Data Version Control git repository that supports an updateable pipeline for extraction, transformation, and loading data into the BioBricks.ai backend at https://biobricks.ai. Use cases include accelerating data science workflows and facilitating the creation of novel data assets by integrating multiple datasets into unified, harmonized resources. In conclusion, BioBricks.ai offers an opportunity to accelerate access and use of public data through a single open platform., Comment: 23 pages, 2 figures

Modern 3D printers allow for the accurate placing of microscopic material voxels, to form complex three-dimensional structures. The advancing capabilities in multi-material printing, coupled with the discovery of printable responsive materials, pave the way for the 3D printing of digital materials, and for the 3D printing of shape-transforming structures, or 4D printing. In this work, we leverage and combine these advancements to develop a versatile methodology for crafting digitized 4D-printed shape-transforming sheets. We 3D print digital responsive sheets, composed of active and passive voxels that are meticulously positioned to form thin structures that can be actuated on demand. This methodology enables us to encode the lateral geometry and the reference curvature, separately and simultaneously. Our approach grants us comprehensive control over the resulting shape, unlocking novel opportunities in synthetic shape-morphing materials. Furthermore, it has the potential to program anomalous mechanical properties and to be extended to multi-material and metamaterial systems., Comment: 12 pages, SI, 5 figures

To mitigate the model dependencies of searches for new narrow resonances at the Large Hadron Collider (LHC), semi-supervised Neural Networks (NNs) can be used. Unlike fully supervised classifiers these models introduce an additional look-elsewhere effect in the process of optimising thresholds on the response distribution. We perform a frequentist study to quantify this effect, in the form of a trials factor. As an example, we consider simulated $Z\gamma$ data to perform narrow resonance searches using semi-supervised NN classifiers. The results from this analysis provide substantiation that the look-elsewhere effect induced by the semi-supervised NN is under control., Comment: 22 pages, 13 figures, with few minor corrections

Bernstein's inequality is a central result in the theory of $D$-modules on smooth varieties. While Bernstein's inequality fails for rings of differential operators on general singularities, recent work of \`{A}lvarez Montaner, Hern\'andez, Jeffries, N\'u\~nez-Betancourt, Teixeira, and Witt establishes Bernstein's inequality for invariants of finite groups in characteristic zero and certain other mild singularities in positive characteristic. Motivated by extending this result to new classes of singular rings, we introduce a ``two-sided'' analogue of the Bernstein-Sato polynomial which we call the sandwich Bernstein-Sato polynomial. We apply this notion to give an effective criterion to verify Bernstein's inequality, and apply this to show that Bernstein's inequality holds for the coordinate ring of $\mathbb{P}^a \times \mathbb{P}^b$ via the Segre embedding. We also establish a number of examples and basic results on sandwich Bernstein-Sato polynomials.

Compressing high-capability Large Language Models (LLMs) has emerged as a favored strategy for resource-efficient inferences. While state-of-the-art (SoTA) compression methods boast impressive advancements in preserving benign task performance, the potential risks of compression in terms of safety and trustworthiness have been largely neglected. This study conducts the first, thorough evaluation of three (3) leading LLMs using five (5) SoTA compression techniques across eight (8) trustworthiness dimensions. Our experiments highlight the intricate interplay between compression and trustworthiness, revealing some interesting patterns. We find that quantization is currently a more effective approach than pruning in achieving efficiency and trustworthiness simultaneously. For instance, a 4-bit quantized model retains the trustworthiness of its original counterpart, but model pruning significantly degrades trustworthiness, even at 50% sparsity. Moreover, employing quantization within a moderate bit range could unexpectedly improve certain trustworthiness dimensions such as ethics and fairness. Conversely, extreme quantization to very low bit levels (3 bits) tends to reduce trustworthiness significantly. This increased risk cannot be uncovered by looking at benign performance alone, in turn, mandating comprehensive trustworthiness evaluation in practice. These findings culminate in practical recommendations for simultaneously achieving high utility, efficiency, and trustworthiness in LLMs. Code and models are available at https://decoding-comp-trust.github.io., Comment: Accepted to ICML'24

Generative adversarial collaborations (GACs) are a form of formal teamwork between groups of scientists with diverging views. The goal of GACs is to identify and ultimately resolve the most important challenges, controversies, and exciting theoretical and empirical debates in a given research field. A GAC team would develop specific, agreed-upon avenues to resolve debates in order to move a field of research forward in a collaborative way. Such adversarial collaborations have many benefits and opportunities but also come with challenges. Here, we use our experience from (1) creating and running the GAC program for the Cognitive Computational Neuroscience (CCN) conference and (2) implementing and leading GACs on particular scientific problems to provide a practical guide for future GAC program organizers and leaders of individual GACs.

Although binary classification is a well-studied problem in computer vision, training reliable classifiers under severe class imbalance remains a challenging problem. Recent work has proposed techniques that mitigate the effects of training under imbalance by modifying the loss functions or optimization methods. While this work has led to significant improvements in the overall accuracy in the multi-class case, we observe that slight changes in hyperparameter values of these methods can result in highly variable performance in terms of Receiver Operating Characteristic (ROC) curves on binary problems with severe imbalance. To reduce the sensitivity to hyperparameter choices and train more general models, we propose training over a family of loss functions, instead of a single loss function. We develop a method for applying Loss Conditional Training (LCT) to an imbalanced classification problem. Extensive experiment results, on both CIFAR and Kaggle competition datasets, show that our method improves model performance and is more robust to hyperparameter choices. Code is available at https://github.com/klieberman/roc_lct.

*This paper is from 2018* In this paper, we try to classify moduli spaces of arrangements of $12$ lines with sextic points. We show that moduli spaces of arrangements of $12$ lines with sextic points can consist of more than two connected components. We also present defining equations of the arrangements whose moduli spaces are not irreducible taking quotients by the complex conjugation by supply some potential Zariski pairs. Through complex conjugation we take quotients and supply some potential Zariski pairs., Comment: *A paper from 2018*. arXiv admin note: text overlap with arXiv:1206.2486

Very large numbers of M-pox cases have, since the start of May 2022, been reported in non-endemic countries leading many to fear that the M-pox Outbreak would rapidly transition into another pandemic, while the COVID-19 pandemic ravages on. Given the similarities of M-pox with COVID-19, we chose to test the performance of COVID-19 models trained on South African twitter data on a hand-labelled M-pox dataset before and after fine-tuning. More than 20k M-pox-related tweets from South Africa were hand-labelled as being either positive, negative or neutral. After fine-tuning these COVID-19 models on the M-pox dataset, the F1-scores increased by more than 8% falling just short of 70%, but still outperforming state-of-the-art models and well-known classification algorithms. An LDA-based topic modelling procedure was used to compare the miss-classified M-pox tweets of the original COVID-19 RoBERTa model with its fine-tuned version, and from this analysis, we were able to draw conclusions on how to build more sophisticated models.