In this work, we introduce Context-Aware MultiModal Learner (CaMML), for tuning large multimodal models (LMMs). CaMML, a lightweight module, is crafted to seamlessly integrate multimodal contextual samples into large models, thereby empowering the model to derive knowledge from analogous, domain-specific, up-to-date information and make grounded inferences. Importantly, CaMML is highly scalable and can efficiently handle lengthy multimodal context examples owing to its hierarchical design. Based on CaMML, we have developed two multimodal models, CaMML-7B and CaMML-13B, that have shown exceptional performance across an array of benchmark datasets for multimodal tasks. Remarkably, CaMML-13B achieves the state-of-the-art performance on over ten widely recognized multimodal benchmark datasets, surpassing LLaVA-1.5 (13B) with a noticeable margin, without integration of any external resources. Moreover, we have conducted extensive ablative studies to inspect the inner workings of CaMML and performed qualitative analyses to showcase its effectiveness in handling real-world challenging cases.
In-context learning (ICL) involves reasoning from given contextual examples. As more modalities comes, this procedure is becoming more challenging as the interleaved input modalities convolutes the understanding process. This is exemplified by the observation that multimodal models often struggle to effectively extrapolate from contextual examples to perform ICL. To address these challenges, we introduce MultiModal In-conteXt Tuning (M$^2$IXT), a lightweight module to enhance the ICL capabilities of multimodal unified models. The proposed M$^2$IXT module perceives an expandable context window to incorporate various labeled examples of multiple modalities (e.g., text, image, and coordinates). It can be prepended to various multimodal unified models (e.g., OFA, Unival, LLaVA) of different architectures and trained via a mixed-tasks strategy to enable rapid few-shot adaption on multiple tasks and datasets. When tuned on as little as 50K multimodal data, M$^2$IXT can boost the few-shot ICL performance significantly (e.g., 18\% relative increase for OFA), and obtained state-of-the-art results across an array of tasks including visual question answering, image captioning, visual grounding, and visual entailment, while being considerably small in terms of model parameters (e.g., $\sim$$20\times$ smaller than Flamingo or MMICL), highlighting the flexibility and effectiveness of M$^2$IXT as a multimodal in-context learner.
Earth system forecasting has traditionally relied on complex physical models that are computationally expensive and require significant domain expertise. In the past decade, the unprecedented increase in spatiotemporal Earth observation data has enabled data-driven forecasting models using deep learning techniques. These models have shown promise for diverse Earth system forecasting tasks but either struggle with handling uncertainty or neglect domain-specific prior knowledge, resulting in averaging possible futures to blurred forecasts or generating physically implausible predictions. To address these limitations, we propose a two-stage pipeline for probabilistic spatiotemporal forecasting: 1) We develop PreDiff, a conditional latent diffusion model capable of probabilistic forecasts. 2) We incorporate an explicit knowledge control mechanism to align forecasts with domain-specific physical constraints. This is achieved by estimating the deviation from imposed constraints at each denoising step and adjusting the transition distribution accordingly. We conduct empirical studies on two datasets: N-body MNIST, a synthetic dataset with chaotic behavior, and SEVIR, a real-world precipitation nowcasting dataset. Specifically, we impose the law of conservation of energy in N-body MNIST and anticipated precipitation intensity in SEVIR. Experiments demonstrate the effectiveness of PreDiff in handling uncertainty, incorporating domain-specific prior knowledge, and generating forecasts that exhibit high operational utility.
Addressing imbalanced or long-tailed data is a major challenge in visual recognition tasks due to disparities between training and testing distributions and issues with data noise. We propose the Wrapped Cauchy Distributed Angular Softmax (WCDAS), a novel softmax function that incorporates data-wise Gaussian-based kernels into the angular correlation between feature representations and classifier weights, effectively mitigating noise and sparse sampling concerns. The class-wise distribution of angular representation becomes a sum of these kernels. Our theoretical analysis reveals that the wrapped Cauchy distribution excels the Gaussian distribution in approximating mixed distributions. Additionally, WCDAS uses trainable concentration parameters to dynamically adjust the compactness and margin of each class. Empirical results confirm label-aware behavior in these parameters and demonstrate WCDAS's superiority over other state-of-the-art softmax-based methods in handling long-tailed visual recognition across multiple benchmark datasets. The code is public available.
Geospatial technologies are becoming increasingly essential in our world for a large range of tasks, such as earth monitoring and natural disaster response. To help improve the applicability and performance of deep learning models on these geospatial tasks, various works have pursued the idea of a geospatial foundation model, i.e., training networks from scratch on a large corpus of remote sensing imagery. However, this approach often requires a significant amount of data and training time to achieve suitable performance, especially when employing large state-of-the-art transformer models. In light of these challenges, we investigate a sustainable approach to building geospatial foundation models. In our investigations, we discover two important factors in the process. First, we find that the selection of pretraining data matters, even within the geospatial domain. We therefore gather a concise yet effective dataset for pretraining. Second, we find that available pretrained models on diverse datasets like ImageNet-22k should not be ignored when building geospatial foundation models, as their representations are still surprisingly effective. Rather, by leveraging their representations, we can build strong models for geospatial applications in a sustainable manner. To this end, we formulate a multi-objective continual pretraining approach for training sustainable geospatial foundation models. We experiment on a wide variety of downstream datasets and tasks, achieving strong performance across the board in comparison to ImageNet baselines and state-of-the-art geospatial pretrained models.
The clarity of microscopic images is vital in biology research and diagnosis. When taking microscopy images at cell or molecule level, mechanical drift occurs and could be difficult and expansive to counter. Such a problem could be overcome by developing an end-to-end deep learning-based workflow capable of predicting in focused microscopic images from out-of-focused counterparts. In our model, we adopt a structure of multi-level U-net, each level connected head-to-tail with corresponding convolution layers from each other. In contrast to the conventional coarse-to-fine model, our model uses the knowledge distilled from the coarse network transferred to the finer network. We evaluate the performance of our model and found our method to be effective and has a better performance by comparing the results with existing models.