Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Suggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works42 from public data
- 646
- Meta-Learning with Task-Adaptive Loss Function for Few-Shot Learning161
This work introduces a new meta-learning framework with a loss function that adapts to each task, named Meta-Learning with Task-Adaptive Loss Function (MeTAL), which demonstrates the effectiveness and the flexibility across various domains, such as few-shot classification and few- shot regression.
- Learning to Forget for Meta-Learning114
This work proposes task-and-layer-wise attenuation on the compromised initialization of model-agnostic meta-learning to reduce its influence and names the method as L2F (Learn to Forget).
- 60
- DAQ: Channel-Wise Distribution-Aware Quantization for Deep Image Super-Resolution Networks55
This work introduces a new ultra-low precision yet effective quantization approach specifically designed for SR, dubbed Distribution-Aware Quantization (DAQ), which manages to greatly reduce the computational and resource costs without the significant sacrifice in SR performance, compared to other quantization methods.
- Scene-Adaptive Video Frame Interpolation via Meta-Learning50
This work shows the benefits of 'test-time adaptation' through simple fine-tuning of a network, then improves its efficiency by incorporating meta-learning, and obtains significant performance gains with only a single gradient update without any additional parameters.
- 45
- 38
- 38
- CLOSER: Towards Better Representation Learning for Few-Shot Class-Incremental Learning33
It is claimed that the closer different classes are, the better for FSCIL, in stark contrast to prior beliefs that the inter-class distance should be maximized.
- LAN: Learning to Adapt Noise for Image Denoising32
A new denoising algorithm is proposed, dubbed Learning-to-Adapt-Noise (LAN), where a learnable noise offset is directly added to a given noisy image to bring a given input noise closer towards the noise distribution a denoising network is trained to handle.
- SANFlow: Semantic-Aware Normalizing Flow for Anomaly Detection28
This work trains NF to map the feature distributions of normal data to different distributions at each location in the given image to enhance the discriminability and highlight the efficacy of the proposed framework in improving the density modeling and thus anomaly detection performance.
- 23
- 20
- MEIL-NeRF: Memory-Efficient Incremental Learning of Neural Radiance Fields17
A Memory-Efficient Incremental Learning algorithm for NeRF (MEIL-NeRF), which takes inspiration from NeRF itself in that a neural network can serve as a memory that returns pixel RGB values, given rays as queries.
- Batch Normalization Tells You Which Filter is Important13
This work shares the observation that the batch normalization (BN) parameters of pre-trained CNNs can be used to estimate the feature distribution of activation outputs, without processing of training data.
- Visual Tracking by Adaptive Continual Meta-Learning11
This work forms the visual tracking problem as a semi-supervised continual learning problem, where only an initial frame is labeled, and dynamically generates the hyperparameters needed for fast initialization and online update to achieve more robustness via adaptively regulating the learning process.
- Searching for Controllable Image Restoration Networks9
A novel framework for controllable image restoration that can effectively restore multiple types and levels of degradation of a corrupted image, named TASNet, which is automatically determined by the neural architecture search algorithm, which optimizes the efficiency-accuracy trade-off of the candidate model architectures.
- Deep Variational Bayesian Modeling of Haze Degradation Process8
This work proposes to take not only a clean image and but also transmission map as latent variables, the posterior distributions of which are parameterized by corresponding neural networks: dehazing and transmission networks, respectively, to account for uncertainties and factors involved in haze degradation.
- 7
- Rethinking RGB Color Representation for Image Restoration Models7
This work augments the RGB representation to hold structural information of local neighborhoods at each pixel while keeping the color information and pixel-grainedness unharmed, resulting in a new representation space, dubbed augmented RGB ($a$RGB) space.
- 7
- Domain Adaptation of Learned Featuresfor Visual Localization6
This work proposes a few-shot domain adaptation framework for learned local features that deals with varying conditions in visual localization, and demonstrates the superior performance over baselines, while using a scarce number of training examples from the target domain.
- 6
- 5
- 5
- Find A Winning Sign: Sign Is All We Need to Win the Lottery3
It is demonstrated that the parameter sign configuration plays a crucial role in conveying useful information for generalization to any randomly initialized network and that any randomly initialized network can be optimized to exhibit low error barriers along the linear path to the sparse network trained by the method by inheriting its sparsity and parameter sign information.
- 3
- 2
- 1
- EGLOCE: Training-Free Energy-Guided Latent Optimization for Concept Erasure1
This work is the first to establish a new paradigm for safe and controllable image generation through dual energy-based guidance during sampling through EGLOCE, a training-free approach that removes unwanted concepts by re-directing noisy latent during inference.
- Training-Free Debiasing of Diffusion Models via CLIP-Guided Denoising Optimization1
Text Embedding Steering Steering Steering (TES) is proposed, a training-free framework that mitigates demographic bias by directly optimizing conditional text embeddings during the diffusion process and outperforms existing training-free baselines in fairness while maintaining competitive image quality.
- BiasEdit: A Training-Free Bias-Detect-and-Edit Framework for Learning Fair Visual Classifiers1
This work introduces BiasEdit, a modular framework that automatically detects bias attributes from the original dataset and edits them to construct a debiased dataset, addressing a fundamental challenge in Web-sourced visual AI, mitigating dataset-induced bias and achieving state-of-the-art debiasing performance even when training data are fully biased.
- SDS-LoRA: Overcoming Anisotropic Gradient Scaling in Low-Rank Adaptation1
A new low-rank parameterization, SDS-LoRA, which Structurally Decouples Singular values from the backward pass is proposed, which improves loss convergence and reduces the gap to full fine-tuning, significantly enhancing adaptation performance.
- Anomaly as Non-Conformity via Training-Free Graph Laplacian Energy Minimization1
A training-free Laplacian graph energy optimization formulation, named ANoCo that scores Anomaly by the cost of Non-Conformity of a query patch to align with a fixed normal manifold, demonstrating the effectiveness of using optimization-induced feature drift as anomaly measure.
- Decoupled Guidance: Disentangling Subject and Context Pathways in Text-to-Image Personalization1
Decoupled Guidance (DeGu) is proposed, a plug-and-play framework that routes subject identity and scene context through two independent guidance streams and introduces a spatial mixing mechanism that dynamically fuses these streams, ensuring each operates within its semantically relevant region without interference.
- Text-to-seed generation: Training-free open-vocabulary seeded semantic segmentation via re-purposing diffusion as text-guided seed generator–
This work proposes Text-to-Seed (T2S), a training-free framework that leverages the text-to-region correspondence of Stable Diffusion to generate attention-based seed points for target categories described by text, and achieves strong performance on standard OVSS benchmarks, demonstrating the effectiveness of combining semantic grounding with seed-driven spatial segmentation.
- –
- –
- –
- Bilinear Coordinate Alignment for Training-Free Task-Vector Transfer–
This work revisits the formation of task vectors and shows that they can be derived as accumulated bilinear interactions between input-side activations and output-side gradients, and proposes BiCo, a training-free framework for transferring task vectors through Bilinear Coordinate alignment.
- –
Publication data from OpenAlex; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.