Despite the success of deep learning in close-set 3D object detection, existing approaches struggle with zero-shot generalization to novel objects and camera configurations. We introduce DetAny3D, a promptable 3D detection foundation model capable of detecting any novel object under arbitrary camera configurations using only monocular inputs. Training a foundation model for 3D detection is fundamentally constrained by the limited availability of annotated 3D data, which motivates DetAny3D to leverage the rich prior knowledge embedded in extensively pre-trained 2D foundation models to compensate for this scarcity. To effectively transfer 2D knowledge to 3D, DetAny3D incorporates two core modules: the 2D Aggregator, which aligns features from different 2D foundation models, and the 3D Interpreter with Zero-Embedding Mapping, which mitigates catastrophic forgetting in 2D-to-3D knowledge transfer. Experimental results validate the strong generalization of our DetAny3D, which not only achieves state-of-the-art performance on unseen categories and novel camera configurations, but also surpasses most competitors on in-domain data.DetAny3D sheds light on the potential of the 3D foundation model for diverse applications in real-world scenarios, e.g., rare object detection in autonomous driving, and demonstrates promise for further exploration of 3D-centric tasks in open-world settings. More visualization results can be found at DetAny3D project page.
Reliable machine learning and statistical analysis rely on diverse, well-distributed training data. However, real-world datasets are often limited in size and exhibit underrepresentation across key subpopulations, leading to biased predictions and reduced performance, particularly in supervised tasks such as classification. To address these challenges, we propose Conditional Data Synthesis Augmentation (CoDSA), a novel framework that leverages generative models, such as diffusion models, to synthesize high-fidelity data for improving model performance across multimodal domains including tabular, textual, and image data. CoDSA generates synthetic samples that faithfully capture the conditional distributions of the original data, with a focus on under-sampled or high-interest regions. Through transfer learning, CoDSA fine-tunes pre-trained generative models to enhance the realism of synthetic data and increase sample density in sparse areas. This process preserves inter-modal relationships, mitigates data imbalance, improves domain adaptation, and boosts generalization. We also introduce a theoretical framework that quantifies the statistical accuracy improvements enabled by CoDSA as a function of synthetic sample volume and targeted region allocation, providing formal guarantees of its effectiveness. Extensive experiments demonstrate that CoDSA consistently outperforms non-adaptive augmentation strategies and state-of-the-art baselines in both supervised and unsupervised settings.
The histopathological images contain a huge amount of information, which can make diagnosis an extremely timeconsuming and tedious task. In this study, we developed a completely automated system to detect regions of interest (ROIs) in whole slide images (WSI) of renal cell carcinoma (RCC), to reduce time analysis and assist pathologists in making more accurate decisions. The proposed approach is based on an efficient texture descriptor named dominant rotated local binary pattern (DRLBP) and color transformation to reveal and exploit the immense texture variability at the microscopic high magnifications level. Thereby, the DRLBPs retain the structural information and utilize the magnitude values in a local neighborhood for more discriminative power. For the classification of the relevant ROIs, feature extraction of WSIs patches was performed on the color channels separately to form the histograms. Next, we used the most frequently occurring patterns as a feature selection step to discard non-informative features. The performances of different classifiers on a set of 1800 kidney cancer patches originating from 12 whole slide images were compared and evaluated. Furthermore, the small size of the image dataset allows to investigate deep learning approach based on transfer learning for image patches classification by using deep features and fine-tuning methods. High recognition accuracy was obtained and the classifiers are efficient, the best precision result was 99.17% achieved with SVM. Moreover, transfer learning models perform well with comparable performance, and the highest precision using ResNet-50 reached 98.50%. The proposed approach results revealed a very efficient image classification and demonstrated efficacy in identifying ROIs. This study presents an automatic system to detect regions of interest relevant to the diagnosis of kidney cancer in whole slide histopathology images.
Information on standing dead trees is important for understanding forest ecosystem functioning and resilience but has been lacking over large geographic regions. Climate change has caused large-scale tree mortality events that can remain undetected due to limited data. In this study, we propose a novel method for segmenting standing dead trees using aerial multispectral orthoimages. Because access to annotated datasets has been a significant problem in forest remote sensing due to the need for forest expertise, we introduce a method for domain transfer by leveraging domain adaptation to learn a transformation from a source domain X to target domain Y. In this Image-to-Image translation task, we aim to utilize available annotations in the target domain by pre-training a segmentation network. When images from a new study site without annotations are introduced (source domain X), these images are transformed into the target domain. Then, transfer learning is applied by inferring the pre-trained network on domain-adapted images. In addition to investigating the feasibility of current domain adaptation approaches for this objective, we propose a novel approach called the Attention-guided Domain Adaptation Network (ADA-Net) with enhanced contrastive learning. Accordingly, the ADA-Net approach provides new state-of-the-art domain adaptation performance levels outperforming existing approaches. We have evaluated the proposed approach using two datasets from Finland and the US. The USA images are converted to the Finland domain, and we show that the synthetic USA2Finland dataset exhibits similar characteristics to the Finland domain images. The software implementation is shared at https://github.com/meteahishali/ADA-Net. The data is publicly available at https://www.kaggle.com/datasets/meteahishali/aerial-imagery-for-standing-dead-tree-segmentation.
Focal cortical dysplasia (FCD) type II is a major cause of drug-resistant epilepsy, often curable only by surgery. Despite its clinical importance, the diagnosis of FCD is very difficult in MRI because of subtle abnormalities, leading to misdiagnosis. This study investigates the use of 3D convolutional neural networks (3D-CNNs) for FCD detection, using a dataset of 170 subjects (85 FCD patients and 85 controls) composed of T1-weighted and FLAIR MRI scans. In particular, it investigates the benefits obtained from cross-modality transfer learning and explainable artificial intelligence (XAI) techniques, in particular Gradient-weighted Class Activation Mapping (Grad-CAM). ResNet architectures (ResNet-18, -34, and -50) were implemented, employing transfer learning strategies that used pre-trained weights from segmentation tasks. Results indicate that transfer learning significantly enhances classification accuracy (up to 80.3%) and interpretability, as measured by a novel Heat-Score metric, which evaluates the model's focus on clinically relevant regions. Improvements in the Heat-Score metric underscore the model's seizure zone localization capabilities, bringing AI predictions and clinical insights closer together. These results highlight the importance of transfer learning, including cross-modality, and XAI in advancing AI-based medical diagnostics, especially for difficult-to-diagnose pathologies such as FCD.
Data-driven approaches such as deep learning can result in predictive models for material properties with exceptional accuracy and efficiency. However, in many problems data is sparse, severely limiting their accuracy and applicability. To improve predictions, techniques such as transfer learning and multi-task learning have been used. The performance of multi-task learning models depends on the strength of the underlying correlations between tasks and the completeness of the dataset. We find that standard multi-task models tend to underperform when trained on sparse datasets with weakly correlated properties. To address this gap, we use data fusion techniques to combine the learned molecular embeddings of various single-task models and trained a multi-task model on this combined embedding. We apply this technique to a widely used benchmark dataset of quantum chemistry data for small molecules as well as a newly compiled sparse dataset of experimental data collected from literature and our own quantum chemistry and thermochemical calculations. The results show that the fused, multi-task models outperform standard multi-task models for sparse datasets and can provide enhanced prediction on data-limited properties compared to single-task models.
The rapid development of universal machine learning interatomic potentials (uMLIPs) has demonstrated the possibility for generalizable learning of the universal potential energy surface. In principle, the accuracy of uMLIPs can be further improved by bridging the model from lower-fidelity datasets to high-fidelity ones. In this work, we analyze the challenge of this transfer learning problem within the CHGNet framework. We show that significant energy scale shifts and poor correlations between GGA and r$^2$SCAN pose challenges to cross-functional data transferability in uMLIPs. By benchmarking different transfer learning approaches on the MP-r$^2$SCAN dataset of 0.24 million structures, we demonstrate the importance of elemental energy referencing in the transfer learning of uMLIPs. By comparing the scaling law with and without the pre-training on a low-fidelity dataset, we show that significant data efficiency can still be achieved through transfer learning, even with a target dataset of sub-million structures. We highlight the importance of proper transfer learning and multi-fidelity learning in creating next-generation uMLIPs on high-fidelity data.
The ability to transfer knowledge and models across dif-ferent datasets and clinical scenarios is of paramount impor-tance in medical imaging. This is especially true for liverlesion segmentation which is crucial for pre-operative plan-ning and treatment follow-up. Despite the progress of deeplearning algorithms using Transformers, automatically seg-menting small hepatic metastases remains a persistent chal-lenge. This can be attributed to the degradation of small struc-tures due to the intrinsic process of feature down-samplinginherent to many architectures as well as class imbalance.While similar challenges have been observed for liver tumorsoriginated from hepatocellular carcinoma, their manifestationin the context of liver metastasis delineation remains under-investigated. Through comprehensive experiments, this paperaims to bridge this gap and to demonstrate the impact of var-ious transfer learning schemes from off-the-shelf datasets toa dataset containing liver metastases only. Our scale-specificevaluation reveals that models trained from scratch or withdomain-specific pre-training demonstrate greater proficiency.
We benchmark foundation models image embeddings for classification and retrieval in e-Commerce, evaluating their suitability for real-world applications. Our study spans embeddings from pre-trained convolutional and transformer models trained via supervised, self-supervised, and text-image contrastive learning. We assess full fine-tuning and transfer learning (top-tuning) on six diverse e-Commerce datasets: fashion, consumer goods, cars, food, and retail. Results show full fine-tuning consistently performs well, while text-image and self-supervised embeddings can match its performance with less training. While supervised embeddings remain stable across architectures, SSL and contrastive embeddings vary significantly, often benefiting from top-tuning. Top-tuning emerges as an efficient alternative to full fine-tuning, reducing computational costs. We also explore cross-tuning, noting its impact depends on dataset characteristics. Our findings offer practical guidelines for embedding selection and fine-tuning strategies, balancing efficiency and performance.
The growing capabilities of large language models (LLMs) present a key challenge of maintaining effective human oversight. Weak-to-strong generalization (W2SG) offers a promising framework for supervising increasingly capable LLMs using weaker ones. Traditional W2SG methods rely on passive learning, where a weak teacher provides noisy demonstrations to train a strong student. This hinders students from employing their knowledge during training and reaching their full potential. In this work, we introduce Alice (pro{A}ctive {l}earning w{i}th tea{c}her's D{e}monstrations), a framework that leverages complementary knowledge between teacher and student to enhance the learning process.We probe the knowledge base of the teacher model by eliciting their uncertainty, and then use these insights together with teachers' responses as demonstrations to guide student models in self-generating improved responses for supervision. In addition, for situations with significant capability gaps between teacher and student models, we introduce cascade Alice, which employs a hierarchical training approach where weak teachers initially supervise intermediate models, who then guide stronger models in sequence. Experimental results demonstrate that our method significantly enhances the W2SG performance, yielding substantial improvements in three key tasks compared to the original W2SG: knowledge-based reasoning (+4.0%), mathematical reasoning (+22.62%), and logical reasoning (+12.11%). This highlights the effectiveness of our new W2SG paradigm that enables more robust knowledge transfer and supervision outcome.
Modeling and forecasting interval-valued time series (ITS) have attracted considerable attention due to their growing presence in various contexts. To the best of our knowledge, there have been no efforts to model large-scale ITS. In this paper, we propose a feature extraction procedure for large-scale ITS, which involves key steps such as auto-segmentation and clustering, and feature transfer learning. This procedure can be seamlessly integrated with any suitable prediction models for forecasting purposes. Specifically, we transform the automatic segmentation and clustering of ITS into the estimation of Toeplitz sparse precision matrices and assignment set. The majorization-minimization algorithm is employed to convert this highly non-convex optimization problem into two subproblems. We derive efficient dynamic programming and alternating direction method to solve these two subproblems alternately and establish their convergence properties. By employing the Joint Recurrence Plot (JRP) to image subsequence and assigning a class label to each cluster, an image dataset is constructed. Then, an appropriate neural network is chosen to train on this image dataset and used to extract features for the next step of forecasting. Real data applications demonstrate that the proposed method can effectively obtain invariant representations of the raw data and enhance forecasting performance.
Foundation models, with a vast number of parameters and pretraining on massive datasets, achieve state-of-the-art performance across various applications. However, efficiently adapting them to downstream tasks with minimal computational overhead remains a challenge. Parameter-Efficient Transfer Learning (PETL) addresses this by fine-tuning only a small subset of parameters while preserving pre-trained knowledge. In this paper, we propose SaS, a novel PETL method that effectively mitigates distributional shifts during fine-tuning. SaS integrates (1) a shared module that captures common statistical characteristics across layers using low-rank projections and (2) a layer-specific module that employs hypernetworks to generate tailored parameters for each layer. This dual design ensures an optimal balance between performance and parameter efficiency while introducing less than 0.05% additional parameters, making it significantly more compact than existing methods. Extensive experiments on diverse downstream tasks, few-shot settings and domain generalization demonstrate that SaS significantly enhances performance while maintaining superior parameter efficiency compared to existing methods, highlighting the importance of capturing both shared and layer-specific information in transfer learning. Code and data are available at https://anonymous.4open.science/r/SaS-PETL-3565.
Diabetic Retinopathy (DR) is a serious and common complication of diabetes, caused by prolonged high blood sugar levels that damage the small retinal blood vessels. If left untreated, DR can progress to retinal vein occlusion and stimulate abnormal blood vessel growth, significantly increasing the risk of blindness. Traditional diabetes diagnosis methods often utilize convolutional neural networks (CNNs) to extract visual features from retinal images, followed by classification algorithms such as decision trees and k-nearest neighbors (KNN) for disease detection. However, these approaches face several challenges, including low accuracy and sensitivity, lengthy machine learning (ML) model training due to high data complexity and volume, and the use of limited datasets for testing and evaluation. This study investigates the application of transfer learning (TL) to enhance ML model performance in DR detection. Key improvements include dimensionality reduction, optimized learning rate adjustments, and advanced parameter tuning algorithms, aimed at increasing efficiency and diagnostic accuracy. The proposed model achieved an overall accuracy of 84% on the testing dataset, outperforming prior studies. The highest class-specific accuracy reached 89%, with a maximum sensitivity of 97% and an F1-score of 92%, demonstrating strong performance in identifying DR cases. These findings suggest that TL-based DR screening is a promising approach for early diagnosis, enabling timely interventions to prevent vision loss and improve patient outcomes.
Electromyography (EMG) signals are used in many applications, including prosthetic hands, assistive suits, and rehabilitation. Recent advances in motion estimation have improved performance, yet challenges remain in cross-subject generalization, electrode shift, and daily variations. When electrode shift occurs, both transfer learning and adversarial domain adaptation improve classification performance by reducing the performance gap to -1\% (eight-class scenario). However, additional data are needed for re-training in transfer learning or for training in adversarial domain adaptation. To address this issue, we investigated a sliding-window normalization (SWN) technique in a real-time prediction scenario. This method combines z-score normalization with a sliding-window approach to reduce the decline in classification performance caused by electrode shift. We validated the effectiveness of SWN using experimental data from a target trajectory tracking task involving the right arm. For three motions classification (rest, flexion, and extension of the elbow) obtained from EMG signals, our offline analysis showed that SWN reduced the differential classification accuracy to -1.0\%, representing a 6.6\% improvement compared to the case without normalization (-7.6\%). Furthermore, when SWN was combined with a strategy that uses a mixture of multiple electrode positions, classification accuracy improved by an additional 2.4\% over the baseline. These results suggest that SWN can effectively reduce the performance degradation caused by electrode shift, thereby enhancing the practicality of EMG-based motion estimation systems.
Simulation-to-Real (Sim2Real) transfer learning, the machine learning technique that efficiently solves a real-world task by leveraging knowledge from computational data, has received increasing attention in materials science as a promising solution to the scarcity of experimental data. We proposed an efficient transfer learning scheme from first-principles calculations to experiments based on the chemistry-informed domain transformation, that integrates the heterogeneous source and target domains by harnessing the underlying physics and chemistry. The proposed method maps the computational data from the simulation space (source domain) into the space of experimental data (target domain). During this process, these qualitatively different domains are efficiently integrated by a couple of prior knowledge of chemistry, (1) the statistical ensemble, and (2) the relationship between source and target quantities. As a proof-of-concept, we predict the catalyst activity for the reverse water-gas shift reaction by using the abundant first-principles data in addition to the experimental data. Through the demonstration, we confirmed that the transfer learning model exhibits positive transfer in accuracy and data efficiency. In particular, a significantly high accuracy was achieved despite using a few (less than ten) target data in domain transformation, whose accuracy is one order of magnitude smaller than that of a full scratch model trained with over 100 target data. This result indicates that the proposed method leverages the high prediction performance with few target data, which helps to save the number of trials in real laboratories.
Proteins move and deform to ensure their biological functions. Despite significant progress in protein structure prediction, approximating conformational ensembles at physiological conditions remains a fundamental open problem. This paper presents a novel perspective on the problem by directly targeting continuous compact representations of protein motions inferred from sparse experimental observations. We develop a task-specific loss function enforcing data symmetries, including scaling and permutation operations. Our method PETIMOT (Protein sEquence and sTructure-based Inference of MOTions) leverages transfer learning from pre-trained protein language models through an SE(3)-equivariant graph neural network. When trained and evaluated on the Protein Data Bank, PETIMOT shows superior performance in time and accuracy, capturing protein dynamics, particularly large/slow conformational changes, compared to state-of-the-art flow-matching approaches and traditional physics-based models.
Learning to generate neural network parameters conditioned on task descriptions and architecture specifications is pivotal for advancing model adaptability and transfer learning. Existing methods especially those based on diffusion models suffer from limited scalability to large architectures, rigidity in handling varying network depths, and disjointed parameter generation that undermines inter-layer coherence. In this work, we propose IGPG (Instruction Guided Parameter Generation), an autoregressive framework that unifies parameter synthesis across diverse tasks and architectures. IGPG leverages a VQ-VAE and an autoregressive model to generate neural network parameters, conditioned on task instructions, dataset, and architecture details. By autoregressively generating neural network weights' tokens, IGPG ensures inter-layer coherence and enables efficient adaptation across models and datasets. Operating at the token level, IGPG effectively captures complex parameter distributions aggregated from a broad spectrum of pretrained models. Extensive experiments on multiple vision datasets demonstrate that IGPG consolidates diverse pretrained models into a single, flexible generative framework. The synthesized parameters achieve competitive or superior performance relative to state-of-the-art methods, especially in terms of scalability and efficiency when applied to large architectures. These results underscore ICPG potential as a powerful tool for pretrained weight retrieval, model selection, and rapid task-specific fine-tuning.
Advanced driver assistance systems require a comprehensive understanding of the driver's mental/physical state and traffic context but existing works often neglect the potential benefits of joint learning between these tasks. This paper proposes MMTL-UniAD, a unified multi-modal multi-task learning framework that simultaneously recognizes driver behavior (e.g., looking around, talking), driver emotion (e.g., anxiety, happiness), vehicle behavior (e.g., parking, turning), and traffic context (e.g., traffic jam, traffic smooth). A key challenge is avoiding negative transfer between tasks, which can impair learning performance. To address this, we introduce two key components into the framework: one is the multi-axis region attention network to extract global context-sensitive features, and the other is the dual-branch multimodal embedding to learn multimodal embeddings from both task-shared and task-specific features. The former uses a multi-attention mechanism to extract task-relevant features, mitigating negative transfer caused by task-unrelated features. The latter employs a dual-branch structure to adaptively adjust task-shared and task-specific parameters, enhancing cross-task knowledge transfer while reducing task conflicts. We assess MMTL-UniAD on the AIDE dataset, using a series of ablation studies, and show that it outperforms state-of-the-art methods across all four tasks. The code is available on https://github.com/Wenzhuo-Liu/MMTL-UniAD.
Existing Speech Language Model (SLM) scaling analysis paints a bleak picture. They predict that SLMs require much more compute and data compared to text, leading some to question the feasibility of training high-quality SLMs. However, modern SLMs are often initialised from pre-trained TextLMs using speech-text interleaving to allow knowledge transfer. This raises the question - Do interleaved SLMs scale more efficiently than textless-SLMs? In this paper we answer a resounding, yes! We conduct scaling analysis of interleaved SLMs by training several dozen and analysing the scaling trends. We see that under this setup SLMs scale more efficiently with compute. Additionally, our results indicate that the scaling-dynamics are significantly different than textless-SLMs, suggesting one should allocate notably more of the compute budget for increasing model size over training tokens. We also study the role of synthetic data and TextLM model families in unlocking this potential. Results suggest, that our scaled up model achieves comparable performance with leading models on speech semantic metrics while using less compute and data than other approaches. We open source models, samples, and data - https://pages.cs.huji.ac.il/adiyoss-lab/sims.
Accurate and quick diagnosis of normal brain tissue Glioma, Meningioma, and Pituitary Tumors is crucial for optimal treatment planning and improved medical results. Magnetic Resonance Imaging (MRI) is widely used as a non-invasive diagnostic tool for detecting brain abnormalities, including tumors. However, manual interpretation of MRI scans is often time-consuming, prone to human error, and dependent on highly specialized expertise. This paper proposes an advanced AI-driven technique to detecting glioma, meningioma, and pituitary brain tumors using YoloV11 and YoloV8 deep learning models. Methods: Using a transfer learning-based fine-tuning approach, we integrate cutting-edge deep learning techniques with medical imaging to classify brain tumors into four categories: No-Tumor, Glioma, Meningioma, and Pituitary Tumors. Results: The study utilizes the publicly accessible CE-MRI Figshare dataset and involves fine-tuning pre-trained models YoloV8 and YoloV11 of 99.49% and 99.56% accuracies; and customized CNN accuracy of 96.98%. The results validate the potential of CNNs in achieving high precision in brain tumor detection and classification, highlighting their transformative role in medical imaging and diagnostics.