SciLT: Long-Tailed Classification in Scientific Image Domains

• URL: http://arxiv.org/abs/2604.03687v1
• Date: Sat, 04 Apr 2026 11:11:47 GMT
• Title: SciLT: Long-Tailed Classification in Scientific Image Domains
• Abstract summary: We investigate scientific long-tailed recognition under a purely visual and parameter-efficient fine-tuning paradigm.<n>Motivated by these findings, we propose SciLT, a framework that exploits multi-level representations through adaptive feature fusion and dual-supervision learning.
• Score: 19.916080605584302
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Long-tailed recognition has benefited from foundation models and fine-tuning paradigms, yet existing studies and benchmarks are mainly confined to natural image domains, where pre-training and fine-tuning data share similar distributions. In contrast, scientific images exhibit distinct visual characteristics and supervision signals, raising questions about the effectiveness of fine-tuning foundation models in such settings. In this work, we investigate scientific long-tailed recognition under a purely visual and parameter-efficient fine-tuning (PEFT) paradigm. Experiments on three scientific benchmarks show that fine-tuning foundation models yields limited gains, and reveal that penultimate-layer features play an important role, particularly for tail classes. Motivated by these findings, we propose SciLT, a framework that exploits multi-level representations through adaptive feature fusion and dual-supervision learning. By jointly leveraging penultimate- and final-layer features, SciLT achieves balanced performance across head and tail classes. Extensive experiments demonstrate that SciLT consistently outperforms existing methods, establishing a strong and practical baseline for scientific long-tailed recognition and providing valuable guidance for adapting foundation models to scientific data with substantial domain shifts.

Related papers

• Modeling Scientific Experiment Scenes: Dataset and Model [35.54149729801777]
  Cross-Modal Dual-Path Generator (CM-DPG) is a model for robust open-vocabulary Scene Graph Generation (SGG)<n>Model enhances object-level semantic representations through joint visual-textual encoding.<n>Experiments on PhysScene and VG150 show CM-DPG achieves competitive performance across multiple evaluation settings.
  arXiv  Detail & Related papers  (2026-08-03T21:20:15Z)
• Faithful, Enriched, and Precise: Benchmarking Natural-Science Illustration Generation by T2I models [49.93401412148884]
  We introduce FEPBench, a benchmark built from carefully selected high-quality scientific illustrations.<n>We evaluate T2I models along three dimensions: instruction faithfulness, reasoning enrichment, and semantic precision.<n>Results show that even state-of-the-art closed-source models, such as GPT Image 2 and Nano Banana Pro, still suffer from text-rendering bottlenecks, limited reasoning enrichment, and difficulty balancing generation richness with precision.
  arXiv  Detail & Related papers  (2026-06-04T09:49:02Z)
• A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers [251.23085679210206]
  Scientific Large Language Models (Sci-LLMs) are transforming how knowledge is represented, integrated, and applied in scientific research.<n>This survey reframes the development of Sci-LLMs as a co-evolution between models and their underlying data substrate.<n>We formulate a unified taxonomy of scientific data and a hierarchical model of scientific knowledge.
  arXiv  Detail & Related papers  (2025-08-28T18:30:52Z)
• SciVid: Cross-Domain Evaluation of Video Models in Scientific Applications [63.92604046592333]
  Video foundation models (FMs) hold considerable promise as general-purpose domain-agnostic approaches.<n>We introduce SciVid, a benchmark comprising five tasks across medical computer vision, animal behavior, and weather forecasting.<n>We adapt six leading ViFMs to SciVid using simple trainable readout modules, establishing strong baselines and demonstrating potential for effective transfer learning.
  arXiv  Detail & Related papers  (2025-07-04T13:48:12Z)
• Massive Supervised Fine-tuning Experiments Reveal How Data, Layer, and Training Factors Shape LLM Alignment Quality [22.105116314320696]
  Supervised fine-tuning (SFT) is a critical step in aligning large language models with human instructions and values.<n>We trained a wide range of base models on a variety of datasets including code generation, mathematical reasoning, and general-domain tasks.<n>We then identified the dataset properties that matter most and examined the layer-wise modifications introduced by SFT.
  arXiv  Detail & Related papers  (2025-06-17T16:13:15Z)
• Large EEG-U-Transformer for Time-Step Level Detection Without Pre-Training [1.3254304182988286]
  We propose a simple U-shaped model to efficiently learn representations by capturing both local and global features.<n>Compared to other window-level classification models, our method directly outputs predictions at the time-step level.<n>Our model won 1st place in the 2025 "seizure detection challenge" organized in the International Conference on Artificial Intelligence in Epilepsy and Other Neurological Disorders.
  arXiv  Detail & Related papers  (2025-04-01T01:33:42Z)
• Towards Effective and Efficient Continual Pre-training of Large Language Models [163.34610964970258]
  Continual pre-training (CPT) has been an important approach for adapting language models to specific domains or tasks. This paper presents a technical report for continually pre-training Llama-3 (8B) It significantly enhances the Chinese language ability and scientific reasoning ability of the backbone model.
  arXiv  Detail & Related papers  (2024-07-26T13:55:21Z)
• Geometric Prior Guided Feature Representation Learning for Long-Tailed Classification [47.09355487357069]
  We propose to leverage the geometric information of the feature distribution of the well-represented head class to guide the model to learn the underlying distribution of the tail class. It aims to make the perturbed features cover the underlying distribution of the tail class as much as possible, thus improving the model's generalization performance in the test domain.
  arXiv  Detail & Related papers  (2024-01-21T09:16:29Z)
• A Novel Benchmark for Few-Shot Semantic Segmentation in the Era of Foundation Models [7.428199805959228]
  Few-shot semantic segmentation (FSS) is a crucial challenge in computer vision.<n>With the emergence of vision foundation models (VFM) as generalist feature extractors, we seek to explore the adaptation of these models for FSS.<n>We propose a novel realistic benchmark with a simple and straightforward adaptation process tailored for this task.
  arXiv  Detail & Related papers  (2024-01-20T19:50:51Z)
• On the Out of Distribution Robustness of Foundation Models in Medical Image Segmentation [47.95611203419802]
  Foundations for vision and language, pre-trained on extensive sets of natural image and text data, have emerged as a promising approach. We compare the generalization performance to unseen domains of various pre-trained models after being fine-tuned on the same in-distribution dataset. We further developed a new Bayesian uncertainty estimation for frozen models and used them as an indicator to characterize the model's performance on out-of-distribution data.
  arXiv  Detail & Related papers  (2023-11-18T14:52:10Z)
• Data-Centric Long-Tailed Image Recognition [49.90107582624604]
  Long-tail models exhibit a strong demand for high-quality data. Data-centric approaches aim to enhance both the quantity and quality of data to improve model performance. There is currently a lack of research into the underlying mechanisms explaining the effectiveness of information augmentation.
  arXiv  Detail & Related papers  (2023-11-03T06:34:37Z)
• Class-Balanced Distillation for Long-Tailed Visual Recognition [100.10293372607222]
  Real-world imagery is often characterized by a significant imbalance of the number of images per class, leading to long-tailed distributions. In this work, we introduce a new framework, by making the key observation that a feature representation learned with instance sampling is far from optimal in a long-tailed setting. Our main contribution is a new training method, that leverages knowledge distillation to enhance feature representations.
  arXiv  Detail & Related papers  (2021-04-12T08:21:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.