FineLIP: Extending CLIP's Reach via Fine-Grained Alignment with Longer Text Inputs

As a pioneering vision-language model, CLIP (Contrastive Language-Image Pre-training) has achieved significant success across various domains and a wide range of downstream vision-language tasks. However, the text encoders in popular CLIP models are limited to processing only 77 text tokens, which constrains their ability to effectively handle longer, detail-rich captions. Additionally, CLIP models often struggle to effectively capture detailed visual and textual information, which hampers their performance on tasks that require fine-grained analysis. To address these limitations, we present a novel approach, \textbf{FineLIP}, that extends the capabilities of CLIP. FineLIP enhances cross-modal text-image mapping by incorporating \textbf{Fine}-grained alignment with \textbf{L}onger text input within the CL\textbf{IP}-style framework. FineLIP first extends the positional embeddings to handle longer text, followed by the dynamic aggregation of local image and text tokens. The aggregated results are then used to enforce fine-grained token-to-token cross-modal alignment. We validate our model on datasets with long, detailed captions across two tasks: zero-shot cross-modal retrieval and text-to-image generation. Quantitative and qualitative experimental results demonstrate the effectiveness of FineLIP, outperforming existing state-of-the-art approaches. Furthermore, comprehensive ablation studies validate the benefits of key design elements within FineLIP.

FedRISE: Rating Induced Sign Election of Gradients for Byzantine Tolerant Federated Aggregation

One of the most common defense strategies against model poisoning in federated learning is to employ a robust aggregator mechanism that makes the training more resilient. Many of the existing Byzantine robust aggregators provide theoretical guarantees and are empirically effective against certain categories of attacks. However, we observe that certain high-strength attacks can subvert the aggregator and collapse the training. In addition, most aggregators require identifying tolerant settings to converge. Impact of attacks becomes more pronounced when the number of Byzantines is near-majority, and becomes harder to evade if the attacker is omniscient with access to data, honest updates and aggregation methods. Motivated by these observations, we develop a robust aggregator called FedRISE for cross-silo FL that is consistent and less susceptible to poisoning updates by an omniscient attacker. The proposed method explicitly determines the optimal direction of each gradient through a sign-voting strategy that uses variance-reduced sparse gradients. We argue that vote weighting based on the cosine similarity of raw gradients is misleading, and we introduce a sign-based gradient valuation function that ignores the gradient magnitude. We compare our method against 8 robust aggregators under 6 poisoning attacks on 3 datasets and architectures. Our results show that existing robust aggregators collapse for at least some attacks under severe settings, while FedRISE demonstrates better robustness because of a stringent gradient inclusion formulation.

A Federated Learning-Friendly Approach for Parameter-Efficient Fine-Tuning of SAM in 3D Segmentation

Adapting foundation models for medical image analysis requires finetuning them on a considerable amount of data because of extreme distribution shifts between natural (source) data used for pretraining and medical (target) data. However, collecting task-specific medical data for such finetuning at a central location raises many privacy concerns. Although Federated learning (FL) provides an effective means for training on private decentralized data, communication costs in federating large foundation models can quickly become a significant bottleneck, impacting the solution's scalability. In this work, we address this problem of efficient communication while ensuring effective learning in FL by combining the strengths of Parameter-Efficient Fine-tuning (PEFT) with FL. Specifically, we study plug-and-play Low-Rank Adapters (LoRA) in a federated manner to adapt the Segment Anything Model (SAM) for 3D medical image segmentation. Unlike prior works that utilize LoRA and finetune the entire decoder, we critically analyze the contribution of each granular component of SAM on finetuning performance. Thus, we identify specific layers to be federated that are very efficient in terms of communication cost while producing on-par accuracy. Our experiments show that retaining the parameters of the SAM model (including most of the decoder) in their original state during adaptation is beneficial because fine-tuning on small datasets tends to distort the inherent capabilities of the underlying foundation model. On Fed-KiTS, our approach decreases communication cost (~48x) compared to full fine-tuning while increasing performance (~6% Dice score) in 3D segmentation tasks. Our approach performs similar to SAMed while achieving ~2.8x reduction in communication and parameters to be finetuned. We further validate our approach with experiments on Fed-IXI and Prostate MRI datasets.

Leveraging Self-Supervised Learning for Fetal Cardiac Planes Classification using Ultrasound Scan Videos

Self-supervised learning (SSL) methods are popular since they can address situations with limited annotated data by directly utilising the underlying data distribution. However, the adoption of such methods is not explored enough in ultrasound (US) imaging, especially for fetal assessment. We investigate the potential of dual-encoder SSL in utilizing unlabelled US video data to improve the performance of challenging downstream Standard Fetal Cardiac Planes (SFCP) classification using limited labelled 2D US images. We study 7 SSL approaches based on reconstruction, contrastive loss, distillation, and information theory and evaluate them extensively on a large private US dataset. Our observations and findings are consolidated from more than 500 downstream training experiments under different settings. Our primary observation shows that for SSL training, the variance of the dataset is more crucial than its size because it allows the model to learn generalisable representations, which improve the performance of downstream tasks. Overall, the BarlowTwins method shows robust performance, irrespective of the training settings and data variations, when used as an initialisation for downstream tasks. Notably, full fine-tuning with 1% of labelled data outperforms ImageNet initialisation by 12% in F1-score and outperforms other SSL initialisations by at least 4% in F1-score, thus making it a promising candidate for transfer learning from US video to image data.