U2A: Unified Unimodal Adaptation for Robust and Efficient Multimodal Learning

Multimodal learning often relies on designing new models and complex training strategies to achieve optimal performance. We present Unified Unimodal Adaptation (U2A), which jointly fine-tunes pretrained unimodal encoders using low-rank adaptation (LoRA) for various multimodal tasks. Our method significantly reduces the number of learnable parameters and eliminates the need for complex training strategies, such as alternating training, gradient modifications, or unimodal fine-tuning. To address missing modalities during both training and testing, we introduce Mask Tokens (MT), which generate missing modality features from available modalities using a single token per modality. This simplifies the process, removing the need for specialized feature estimation or prompt-tuning methods. Our evaluation demonstrates that U2A matches or outperforms state-of-the-art methods in both complete and missing modality settings, showcasing strong performance and robustness across various modalities, tasks, and datasets. We also analyze and report the effectiveness of Mask Tokens in different missing modality scenarios. Overall, our method provides a robust, flexible, and efficient solution for multimodal learning, with minimal computational overhead.

MMP: Towards Robust Multi-Modal Learning with Masked Modality Projection

Multimodal learning seeks to combine data from multiple input sources to enhance the performance of different downstream tasks. In real-world scenarios, performance can degrade substantially if some input modalities are missing. Existing methods that can handle missing modalities involve custom training or adaptation steps for each input modality combination. These approaches are either tied to specific modalities or become computationally expensive as the number of input modalities increases. In this paper, we propose Masked Modality Projection (MMP), a method designed to train a single model that is robust to any missing modality scenario. We achieve this by randomly masking a subset of modalities during training and learning to project available input modalities to estimate the tokens for the masked modalities. This approach enables the model to effectively learn to leverage the information from the available modalities to compensate for the missing ones, enhancing missing modality robustness. We conduct a series of experiments with various baseline models and datasets to assess the effectiveness of this strategy. Experiments demonstrate that our approach improves robustness to different missing modality scenarios, outperforming existing methods designed for missing modalities or specific modality combinations.

Compressing GANs using Knowledge Distillation

Generative Adversarial Networks (GANs) have been used in several machine learning tasks such as domain transfer, super resolution, and synthetic data generation. State-of-the-art GANs often use tens of millions of parameters, making them expensive to deploy for applications in low SWAP (size, weight, and power) hardware, such as mobile devices, and for applications with real time capabilities. There has been no work found to reduce the number of parameters used in GANs. Therefore, we propose a method to compress GANs using knowledge distillation techniques, in which a smaller "student" GAN learns to mimic a larger "teacher" GAN. We show that the distillation methods used on MNIST, CIFAR-10, and Celeb-A datasets can compress teacher GANs at ratios of 1669:1, 58:1, and 87:1, respectively, while retaining the quality of the generated image. From our experiments, we observe a qualitative limit for GAN's compression. Moreover, we observe that, with a fixed parameter budget, compressed GANs outperform GANs trained using standard training methods. We conjecture that this is partially owing to the optimization landscape of over-parameterized GANs which allows efficient training using alternating gradient descent. Thus, training an over-parameterized GAN followed by our proposed compression scheme provides a high quality generative model with a small number of parameters.

Understanding the Energy and Precision Requirements for Online Learning

It is well-known that the precision of data, hyperparameters, and internal representations employed in learning systems directly impacts its energy, throughput, and latency. The precision requirements for the training algorithm are also important for systems that learn on-the-fly. Prior work has shown that the data and hyperparameters can be quantized heavily without incurring much penalty in classification accuracy when compared to floating point implementations. These works suffer from two key limitations. First, they assume uniform precision for the classifier and for the training algorithm and thus miss out on the opportunity to further reduce precision. Second, prior works are empirical studies. In this article, we overcome both these limitations by deriving analytical lower bounds on the precision requirements of the commonly employed stochastic gradient descent (SGD) on-line learning algorithm in the specific context of a support vector machine (SVM). Lower bounds on the data precision are derived in terms of the the desired classification accuracy and precision of the hyperparameters used in the classifier. Additionally, lower bounds on the hyperparameter precision in the SGD training algorithm are obtained. These bounds are validated using both synthetic and the UCI breast cancer dataset. Additionally, the impact of these precisions on the energy consumption of a fixed-point SVM with on-line training is studied.