A unified backbone-expert framework with a common convolutional state-space backbone and two specialized interfaces for automatic modulation recognition, confirming the benefit of expert-interface decoupling over one-size-fits-all architectures.
Abstract
Automatic modulation recognition (AMR) faces distinct representation bottlenecks under varying observation lengths, where a single model architecture often fails to excel. To address this, we propose a unified backbone-expert framework with a common convolutional state-space backbone and two specialized interfaces. For short sequences, we inject explicit lag-aware complex-plane descriptors as relation tokens before encoding to compensate for information loss. For long sequences, we design a gated multi-scale residual refinement module to correct the feature map, combined with a fixed-averaging classifier collaboration to harness complementary evidence. Our framework achieves overall average accuracies of 67.28 \pm 0.14% on RML2016.10b and 87.19 \pm 0.77% on HisarMod2019 (mean \pm sample standard deviation over three runs), respectively. The framework's efficacy is further validated through three-seed ablations, native-length cross-configuration tests, and controlled window studies, confirming the benefit of expert-interface decoupling over one-size-fits-all architectures.
ARMDIL is an ensemble that uses a multimodal large language model (MLLM) agent to dynamically route each image to the most suitable vision backbone, drastically improves adaptability by allowing new information to be integrated via simple prompt modifications, while enhancing interpretability through natural language reasoning traces.
Daniel A. Perkins, J. Squires, Janou Milligan et al.· 0 citations
Implicit Neural Representations (INRs) encode signals as the weights of a coordinate-based neural network and have recently been proposed as an alternative domain for downstream learning. While promising, classification directly in weight space remains challenging due to the high dimensionality and complex structure of INR parameters. Furthermore, the way discriminative information is distributed across INR weights remains poorly understood. We propose a hierarchical Mixture-of-Experts (HMoE) Transformer that processes INR weights using conditional computation aligned with the structure of the underlying implicit network. Coupled with a meta-learning framework that shapes INR parameters for downstream tasks, our model achieves state-of-the-art accuracy across standard benchmarks, ranging from low-resolution datasets to high-resolution ImageNet-1K. To gain insight into how INRs encode discriminative information, we develop weight-space attribution and pruning methods that identify parameters most relevant for classification. These analyses reveal how class-specific structure emerges within INR layers and support the suitability of MoE architectures for weight-space learning. Our approach advances both the performance and interpretability of weight-space classifiers.
Stanislaw Janik, Michal Byra· arXiv.org· 0 citations
Token aggregation converts token-level representations into fixed-dimensional sample representations, but most pooling methods operate only in the original token space. We introduce Frequency-Domain Latent-attention Gated Pooling (FLaG), a plug-in aggregation module that re-expresses encoder outputs in the Fourier domain before final pooling. FLaG represents the nonredundant rFFT spectrum through concatenated real and imaginary components, summarizes spectral tokens with learnable latent queries, derives a sample-conditioned channel gate, and reconstructs modulated token representations for downstream aggregation. We evaluate the same architecture across ESM2-based antimicrobial peptide (AMP) activity prediction, ResNet18 image classification on CIFAR-10 and CIFAR-100, and three RoBERTa-based language tasks. FLaG achieves the best macro-averaged Spearman correlation coefficient, RMSE, and Recall@50 across four AMP backbone-species settings and the highest top-1 accuracy on CIFAR 10. It also achieves the best mean results on five of seven language metrics, although mean pooling remains strongest on STSBenchmark. AMP-side mechanistic analyses reveal low-frequency prediction sensitivity across most encoder layers, with increased relative high-frequency sensitivity in the final layer, and pronounced peptide-specific positional responses. The residual gate broadly amplifies spectral channels while preserving the low-frequency-dominated energy profile, whereas latent cross-attention exhibits sample- and species-specific spectral allocation. Overall, FLaG provides a transferable frequency-domain aggregation bias across protein, visual, and textual representations, with benefits that depend on the backbone and downstream task. Supplementary materials, source code, and data are available at https://www.healthinformaticslab.org/supp/ and https://github.com/Kewei2023/AMPCliff/tree/FLaG.
Ke-Wei Li, Rong Zhang, Xuelin Wang et al.· 0 citations
Recent Visual-Language Models (VLMs) have enhanced the capabilities of pre-trained LLMs by adding vision tokens alongside text, with approaches like LLaVA showing impressive results. However, the computational burden of processing up to 576 or 729 visual tokens makes edge deployment challenging. While various token pruning techniques require retraining, some are training-free and thus can easily adapt to architecture changes. We introduce ClustRS, a two-part, training-free algorithm for robust token pruning. Its first component is an attention-weighted, clustering algorithm that selects representative tokens from each semantic cluster. The second component, Residual Shrinkage, is a one-pass denoising step on the selected tokens. These training-free lightweight steps make LLaVA ready for real-world data, improving robustness to a wide range of image-noise types and intensities. Experimental results on the ScienceQA-IMG and MM-VET benchmarks show our method outperforms attention- and diversity-based methods by up to 20\% under extreme noise and token conditions (reducing tokens by 97\%, down to 16 tokens) on LLaVA 1.5 7b and achieves exceptional results on LLaVA-OneVision, where we match baseline performance with fewer than one-third of their tokens under mild noise conditions. Our study demonstrates a simple yet powerful alternative to both score-only and diversity-only pruning rules, paving the way for compute-efficient and noise-resilient VLM deployment.
Baptiste Rossigneux, Inna Kucher, Vincent Lorrain et al.· 0 citations
Recent vision backbones increasingly rely on sophisticated modules, whereas long-used operators such as residual connections, activations, and normalization layers remain less explored for lightweight redesign. This paper revisits these operators and proposes three operator-level redesigns: Subtractive Residual Connection (SRC), Learnable Gating Response Function (LGRF), and statistics-discrepancy-guided Dynamic Dual Normalization (DDN). SRC changes shallow residual fusion from addition to subtraction to suppress redundant responses and induce attention-like response focusing without an explicit attention branch; LGRF extends fixed gating activations to channel-wise learnable response curves; and DDN generates sample-level LayerNorm (LN)–BatchNorm (BN) fusion weights from input statistics and LN-BN discrepancy. On ImageNet-1K, SRC improves ResNet models without extra parameters or floating-point operations (FLOPs). Applying all three proposed methods to MambaOut-Femto improves Top-1 accuracy by 1.01 percentage points with only 0.06 M additional parameters and 0.01 GFLOPs. Ablations on ImageNet-100 and CIFAR-10/100 support effectiveness and stability, while DeepWeeds validation further supports the practical value of all three operators for weed recognition under complex natural backgrounds. Gradient-weighted Class Activation Mapping (Grad-CAM), feature response maps, learned response curves, and dynamic LN-BN weights support module interpretation. Overall, revisiting long-used fundamental operators remains valuable for efficient visual recognition.
MLLM-based segmentation faces a core segmentation trilemma: high segmentation performance, preserved dialogue ability, and fast inference. Embedding-prediction methods may disrupt language modeling through pixel-level objectives, whereas next-token generation is inefficient for dense masks. We propose All-Mask Prediction, decoupling autoregressive dialogue from non-autoregressive mask prediction. Its binary instantiation, STAMP (Simultaneous Textual All-Mask Prediction), emits an in-vocabularytrigger, fuses image-aligned mask tokens with corresponding patch features, and uses hybrid attention to classify all tokens as foreground or background in one pass. It thereby combines strong referring and reasoning segmentation with preserved multimodal ability and efficient inference. However, binary masks cannot retain multiple semantic or instance identities without repeated target-specific predictions. We therefore propose Structured All-Mask Prediction and develop STAMPlus. It generates a target list with explicit IDs and optional boxes, binds these IDs to a shared multi-class mask space, and jointly predicts all targets in one non-autoregressive pass. A single unified checkpoint retains STAMP's referring and reasoning capabilities while extending to open-vocabulary semantic, instance-aware, and remote-sensing small-target segmentation, where high-resolution mask-token scaling preserves finer spatial evidence. Across these settings, STAMPlus achieves state-of-the-art segmentation performance, preserves general multimodal instruction following, and reduces 12-category latency from 13.50s for repeated STAMP inference to 5.16s. Further analyses show that accurate target cues improve segmentation and learned spatial grounding benefits look-twice reasoning. Overall, STAMPlus resolves the trilemma beyond single-target prediction.
Jiazhen Liu, Mingkuan Feng, Long Chen· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
