我们提出了Clip-Lite，一种通过与文本注释的特征对齐方式进行视觉表示学习的信息有效方法。与先前提出的剪辑模型相比，剪辑液在优化其对比学学习目标期间只需要一个负图像文本样本对。我们通过利用信息有效的较低限制来实现这一点，以最大化两个输入模态之间的相互信息。这允许剪辑Lite培训，在获得比夹子的更好的性能的同时具有显着减少的数据和批量尺寸。我们通过在Coco-Tablions数据集上预先绘制来评估剪贴画并对其他数据集进行测试传输。 Clip-Lite在Pascal VOC分类上获得+ 15.4％的映射绝对增益，并在ImageNet上获得A + 22.1％的前1个精度增益，同时与其他更复杂，文本监督模型相当或优越。 Clip-Lite还优于剪辑图像和文本检索，零拍分类和视觉接地。最后，通过在表示学习期间执行显式图像文本对齐，我们显示Clip-Lite可以利用语言语义来鼓励可以在下游任务中使用的无偏见的视觉表示。

在这项工作中，我们提出了相互信息最大化知识蒸馏（MIMKD）。我们的方法使用对比目标来同时估计，并最大化教师和学生网络之间的本地和全球特征表示的相互信息的下限。我们通过广泛的实验证明，这可以通过将知识从更加性能但计算昂贵的模型转移来改善低容量模型的性能。这可用于产生更好的模型，可以在具有低计算资源的设备上运行。我们的方法灵活，我们可以将具有任意网络架构的教师蒸馏到任意学生网络。我们的经验结果表明，MIMKD优于各种学生教师对的竞争方法，具有不同的架构，以及学生网络的容量极低。我们能够通过从Reset-50蒸馏出来的知识，从基线精度为Shufflenetv2获得74.55％的精度。在Imagenet上，我们使用Reset-34教师网络将Reset-18网络从68.88％提高到70.32％的准确度（1.44％+）。

Implicit Neural Representations (INR) have recently shown to be powerful tool for high-quality video compression. However, existing works are limiting as they do not explicitly exploit the temporal redundancy in videos, leading to a long encoding time. Additionally, these methods have fixed architectures which do not scale to longer videos or higher resolutions. To address these issues, we propose NIRVANA, which treats videos as groups of frames and fits separate networks to each group performing patch-wise prediction. This design shares computation within each group, in the spatial and temporal dimensions, resulting in reduced encoding time of the video. The video representation is modeled autoregressively, with networks fit on a current group initialized using weights from the previous group's model. To further enhance efficiency, we perform quantization of the network parameters during training, requiring no post-hoc pruning or quantization. When compared with previous works on the benchmark UVG dataset, NIRVANA improves encoding quality from 37.36 to 37.70 (in terms of PSNR) and the encoding speed by 12X, while maintaining the same compression rate. In contrast to prior video INR works which struggle with larger resolution and longer videos, we show that our algorithm is highly flexible and scales naturally due to its patch-wise and autoregressive designs. Moreover, our method achieves variable bitrate compression by adapting to videos with varying inter-frame motion. NIRVANA achieves 6X decoding speed and scales well with more GPUs, making it practical for various deployment scenarios.

The ability to jointly learn from multiple modalities, such as text, audio, and visual data, is a defining feature of intelligent systems. While there have been promising advances in designing neural networks to harness multimodal data, the enormous success of data augmentation currently remains limited to single-modality tasks like image classification. Indeed, it is particularly difficult to augment each modality while preserving the overall semantic structure of the data; for example, a caption may no longer be a good description of an image after standard augmentations have been applied, such as translation. Moreover, it is challenging to specify reasonable transformations that are not tailored to a particular modality. In this paper, we introduce LeMDA, Learning Multimodal Data Augmentation, an easy-to-use method that automatically learns to jointly augment multimodal data in feature space, with no constraints on the identities of the modalities or the relationship between modalities. We show that LeMDA can (1) profoundly improve the performance of multimodal deep learning architectures, (2) apply to combinations of modalities that have not been previously considered, and (3) achieve state-of-the-art results on a wide range of applications comprised of image, text, and tabular data.

In post-covid19 world, radio frequency (RF)-based non-contact methods, e.g., software-defined radios (SDR)-based methods have emerged as promising candidates for intelligent remote sensing of human vitals, and could help in containment of contagious viruses like covid19. To this end, this work utilizes the universal software radio peripherals (USRP)-based SDRs along with classical machine learning (ML) methods to design a non-contact method to monitor different breathing abnormalities. Under our proposed method, a subject rests his/her hand on a table in between the transmit and receive antennas, while an orthogonal frequency division multiplexing (OFDM) signal passes through the hand. Subsequently, the receiver extracts the channel frequency response (basically, fine-grained wireless channel state information), and feeds it to various ML algorithms which eventually classify between different breathing abnormalities. Among all classifiers, linear SVM classifier resulted in a maximum accuracy of 88.1\%. To train the ML classifiers in a supervised manner, data was collected by doing real-time experiments on 4 subjects in a lab environment. For label generation purpose, the breathing of the subjects was classified into three classes: normal, fast, and slow breathing. Furthermore, in addition to our proposed method (where only a hand is exposed to RF signals), we also implemented and tested the state-of-the-art method (where full chest is exposed to RF radiation). The performance comparison of the two methods reveals a trade-off, i.e., the accuracy of our proposed method is slightly inferior but our method results in minimal body exposure to RF radiation, compared to the benchmark method.

Sentence simplification aims at making the structure of text easier to read and understand while maintaining its original meaning. This can be helpful for people with disabilities, new language learners, or those with low literacy. Simplification often involves removing difficult words and rephrasing the sentence. Previous research have focused on tackling this task by either using external linguistic databases for simplification or by using control tokens for desired fine-tuning of sentences. However, in this paper we purely use pre-trained transformer models. We experiment with a combination of GPT-2 and BERT models, achieving the best SARI score of 46.80 on the Mechanical Turk dataset, which is significantly better than previous state-of-the-art results. The code can be found at https://github.com/amanbasu/sentence-simplification.

This work presents a physics-informed deep learning-based super-resolution framework to enhance the spatio-temporal resolution of the solution of time-dependent partial differential equations (PDE). Prior works on deep learning-based super-resolution models have shown promise in accelerating engineering design by reducing the computational expense of traditional numerical schemes. However, these models heavily rely on the availability of high-resolution (HR) labeled data needed during training. In this work, we propose a physics-informed deep learning-based framework to enhance the spatial and temporal resolution of coarse-scale (both in space and time) PDE solutions without requiring any HR data. The framework consists of two trainable modules independently super-resolving the PDE solution, first in spatial and then in temporal direction. The physics based losses are implemented in a novel way to ensure tight coupling between the spatio-temporally refined outputs at different times and improve framework accuracy. We analyze the capability of the developed framework by investigating its performance on an elastodynamics problem. It is observed that the proposed framework can successfully super-resolve (both in space and time) the low-resolution PDE solutions while satisfying physics-based constraints and yielding high accuracy. Furthermore, the analysis and obtained speed-up show that the proposed framework is well-suited for integration with traditional numerical methods to reduce computational complexity during engineering design.

Vision Transformers (ViTs) have gained significant popularity in recent years and have proliferated into many applications. However, it is not well explored how varied their behavior is under different learning paradigms. We compare ViTs trained through different methods of supervision, and show that they learn a diverse range of behaviors in terms of their attention, representations, and downstream performance. We also discover ViT behaviors that are consistent across supervision, including the emergence of Offset Local Attention Heads. These are self-attention heads that attend to a token adjacent to the current token with a fixed directional offset, a phenomenon that to the best of our knowledge has not been highlighted in any prior work. Our analysis shows that ViTs are highly flexible and learn to process local and global information in different orders depending on their training method. We find that contrastive self-supervised methods learn features that are competitive with explicitly supervised features, and they can even be superior for part-level tasks. We also find that the representations of reconstruction-based models show non-trivial similarity to contrastive self-supervised models. Finally, we show how the "best" layer for a given task varies by both supervision method and task, further demonstrating the differing order of information processing in ViTs.

Modern deep learning models are over-parameterized, where the optimization setup strongly affects the generalization performance. A key element of reliable optimization for these systems is the modification of the loss function. Sharpness-Aware Minimization (SAM) modifies the underlying loss function to guide descent methods towards flatter minima, which arguably have better generalization abilities. In this paper, we focus on a variant of SAM known as mSAM, which, during training, averages the updates generated by adversarial perturbations across several disjoint shards of a mini-batch. Recent work suggests that mSAM can outperform SAM in terms of test accuracy. However, a comprehensive empirical study of mSAM is missing from the literature -- previous results have mostly been limited to specific architectures and datasets. To that end, this paper presents a thorough empirical evaluation of mSAM on various tasks and datasets. We provide a flexible implementation of mSAM and compare the generalization performance of mSAM to the performance of SAM and vanilla training on different image classification and natural language processing tasks. We also conduct careful experiments to understand the computational cost of training with mSAM, its sensitivity to hyperparameters and its correlation with the flatness of the loss landscape. Our analysis reveals that mSAM yields superior generalization performance and flatter minima, compared to SAM, across a wide range of tasks without significantly increasing computational costs.

Accurate segmentation of live cell images has broad applications in clinical and research contexts. Deep learning methods have been able to perform cell segmentations with high accuracy; however developing machine learning models to do this requires access to high fidelity images of live cells. This is often not available due to resource constraints like limited accessibility to high performance microscopes or due to the nature of the studied organisms. Segmentation on low resolution images of live cells is a difficult task. This paper proposes a method to perform live cell segmentation with low resolution images by performing super-resolution as a pre-processing step in the segmentation pipeline.