Fine-tuning pre-trained models has become a cornerstone of modern machine learning, allowing practitioners to achieve high performance with limited labeled data. In surgical video analysis, where expert annotations are especially time-consuming and costly, identifying the most suitable pre-trained model for a downstream task is both critical and challenging. Source-independent transferability estimation (SITE) offers a solution by predicting how well a model will fine-tune on target data using only its embeddings or outputs, without requiring full retraining. In this work, we formalize SITE for surgical phase recognition and provide the first comprehensive benchmark of three representative metrics, LogME, H-Score, and TransRate, on two diverse datasets (RAMIE and AutoLaparo). Our results show that LogME, particularly when aggregated by the minimum per-subset score, aligns most closely with fine-tuning accuracy; H-Score yields only weak predictive power; and TransRate often inverses true model rankings. Ablation studies show that when candidate models have similar performances, transferability estimates lose discriminative power, emphasizing the importance of maintaining model diversity or using additional validation. We conclude with practical guidelines for model selection and outline future directions toward domain-specific metrics, theoretical foundations, and interactive benchmarking tools.

The data used during training in any given application space is directly tied to the performance of the system once deployed. While there are many other factors that go into producing high performance models within machine learning, there is no doubt that the data used to train a system provides the foundation from which to build. One of the underlying rule of thumb heuristics used within the machine learning space is that more data leads to better models, but there is no easy answer for the question, "How much data is needed?" This work examines a modulation classification problem in the Radio Frequency domain space, attempting to answer the question of how much training data is required to achieve a desired level of performance, but the procedure readily applies to classification problems across modalities. The ultimate goal is determining an approach that requires the least amount of data collection to better inform a more thorough collection effort to achieve the desired performance metric. While this approach will require an initial dataset that is germane to the problem space to act as a \textit{target} dataset on which metrics are extracted, the goal is to allow for the initial data to be orders of magnitude smaller than what is required for delivering a system that achieves the desired performance. An additional benefit of the techniques presented here is that the quality of different datasets can be numerically evaluated and tied together with the quantity of data, and ultimately, the performance of the architecture in the problem domain.

We consider transferability estimation, the problem of estimating how well deep learning models transfer from a source to a target task. We focus on regression tasks, which received little previous attention, and propose two simple and computationally efficient approaches that estimate transferability based on the negative regularized mean squared error of a linear regression model. We prove novel theoretical results connecting our approaches to the actual transferability of the optimal target models obtained from the transfer learning process. Despite their simplicity, our approaches significantly outperform existing state-of-the-art regression transferability estimators in both accuracy and efficiency. On two large-scale keypoint regression benchmarks, our approaches yield 12% to 36% better results on average while being at least 27% faster than previous state-of-the-art methods.

This paper investigates an under-explored but important problem: given a collection of pre-trained neural networks, predicting their performance on each multi-modal task without fine-tuning them, such as image recognition, referring, captioning, visual question answering, and text question answering. A brute-force approach is to finetune all models on all target datasets, bringing high computational costs. Although recent-advanced approaches employed lightweight metrics to measure models' transferability,they often depend heavily on the prior knowledge of a single task, making them inapplicable in a multi-modal multi-task scenario. To tackle this issue, we propose an efficient multi-task model selector (EMMS), which employs large-scale foundation models to transform diverse label formats such as categories, texts, and bounding boxes of different downstream tasks into a unified noisy label embedding. EMMS can estimate a model's transferability through a simple weighted linear regression, which can be efficiently solved by an alternating minimization algorithm with a convergence guarantee. Extensive experiments on 5 downstream tasks with 24 datasets show that EMMS is fast, effective, and generic enough to assess the transferability of pre-trained models, making it the first model selection method in the multi-task scenario. For instance, compared with the state-of-the-art method LogME enhanced by our label embeddings, EMMS achieves 9.0\%, 26.3\%, 20.1\%, 54.8\%, 12.2\% performance gain on image recognition, referring, captioning, visual question answering, and text question answering, while bringing 5.13x, 6.29x, 3.59x, 6.19x, and 5.66x speedup in wall-clock time, respectively. The code is available at https://github.com/OpenGVLab/Multitask-Model-Selector.

With the increase in availability of large pre-trained language models (LMs) in Natural Language Processing (NLP), it becomes critical to assess their fit for a specific target task a priori - as fine-tuning the entire space of available LMs is computationally prohibitive and unsustainable. However, encoder transferability estimation has received little to no attention in NLP. In this paper, we propose to generate quantitative evidence to predict which LM, out of a pool of models, will perform best on a target task without having to fine-tune all candidates. We provide a comprehensive study on LM ranking for 10 NLP tasks spanning the two fundamental problem types of classification and structured prediction. We adopt the state-of-the-art Logarithm of Maximum Evidence (LogME) measure from Computer Vision (CV) and find that it positively correlates with final LM performance in 94% of the setups. In the first study of its kind, we further compare transferability measures with the de facto standard of human practitioner ranking, finding that evidence from quantitative metrics is more robust than pure intuition and can help identify unexpected LM candidates.

Model hubs with many pre-trained models (PTMs) have become a cornerstone of deep learning. Although built at a high cost, they remain \emph{under-exploited} -- practitioners usually pick one PTM from the provided model hub by popularity and then fine-tune the PTM to solve the target task. This na\"ive but common practice poses two obstacles to full exploitation of pre-trained model hubs: first, the PTM selection by popularity has no optimality guarantee, and second, only one PTM is used while the remaining PTMs are ignored. An alternative might be to consider all possible combinations of PTMs and extensively fine-tune each combination, but this would not only be prohibitive computationally but may also lead to statistical over-fitting. In this paper, we propose a new paradigm for exploiting model hubs that is intermediate between these extremes. The paradigm is characterized by two aspects: (1) We use an evidence maximization procedure to estimate the maximum value of label evidence given features extracted by pre-trained models. This procedure can rank all the PTMs in a model hub for various types of PTMs and tasks \emph{before fine-tuning}. (2) The best ranked PTM can either be fine-tuned and deployed if we have no preference for the model's architecture or the target PTM can be tuned by the top $K$ ranked PTMs via a Bayesian procedure that we propose. This procedure, which we refer to as \emph{B-Tuning}, not only improves upon specialized methods designed for tuning homogeneous PTMs, but also applies to the challenging problem of tuning heterogeneous PTMs where it yields a new level of benchmark performance.

This paper studies task adaptive pre-trained model selection, an \emph{underexplored} problem of assessing pre-trained models so that models suitable for the task can be selected from the model zoo without fine-tuning. A pilot work~\cite{nguyen_leep:_2020} addressed the problem in transferring supervised pre-trained models to classification tasks, but it cannot handle emerging unsupervised pre-trained models or regression tasks. In pursuit of a practical assessment method, we propose to estimate the maximum evidence (marginalized likelihood) of labels given features extracted by pre-trained models. The maximum evidence is \emph{less prone to over-fitting} than the likelihood, and its \emph{expensive computation can be dramatically reduced} by our carefully designed algorithm. The Logarithm of Maximum Evidence (LogME) can be used to assess pre-trained models for transfer learning: a pre-trained model with high LogME is likely to have good transfer performance. LogME is fast, accurate, and general, characterizing it as \emph{the first practical assessment method for transfer learning}. Compared to brute-force fine-tuning, LogME brings over $3000\times$ speedup in wall-clock time. It outperforms prior methods by a large margin in their setting and is applicable to new settings that prior methods cannot deal with. It is general enough to diverse pre-trained models (supervised pre-trained and unsupervised pre-trained), downstream tasks (classification and regression), and modalities (vision and language). Code is at \url{https://github.com/thuml/LogME}.