Spatial-spectral radio map estimation (RME) from sparsely deployed sensors can be viewed as a tensor learning problem. Among tensor models, the block-term decomposition (BTD) model is especially effective for RME, enabling accurate recovery of the radio map as well as emitter-level spatial loss field (SLF) and power spectral density (PSD). However, selecting the appropriate tensor and block-term ranks in the BTD model is crucial yet challenging, as the ranks indicate the number and properties of emitters, which are typically unknown in practice. Inappropriate settings of these ranks leads to overfitting and underfitting of the model. Existing optimization-based BTD learning methods either unrealistically assume that these ranks are known or require parameter tuning. To address these issues, we propose a Bayesian BTD model that incorporates sparsitypromoting priors on the factor matrices for rank learning, eliminating the need for parameter tuning while learning all other parameters. Extensive numerical experiments showcase the effectiveness of the proposed method.

Invasive brain-computer interfaces with Electrocorticography (ECoG) have shown promise for high-performance speech decoding in medical applications, but less damaging methods like intracranial stereo-electroencephalography (sEEG) remain underexplored. With rapid advances in representation learning, leveraging abundant recordings to enhance speech decoding is increasingly attractive. However, popular methods often pre-train temporal models based on brain-level tokens, overlooking that brain activities in different regions are highly desynchronized during tasks. Alternatively, they pre-train spatial-temporal models based on channel-level tokens but fail to evaluate them on challenging tasks like speech decoding, which requires intricate processing in specific language-related areas. To address this issue, we collected a well-annotated Chinese word-reading sEEG dataset targeting language-related brain networks from 12 subjects. Using this benchmark, we developed the Du-IN model, which extracts contextual embeddings based on region-level tokens through discrete codex-guided mask modeling. Our model achieves state-of-the-art performance on the 61-word classification task, surpassing all baselines. Model comparisons and ablation studies reveal that our design choices, including (\romannumeral1) temporal modeling based on region-level tokens by utilizing 1D depthwise convolution to fuse channels in the ventral sensorimotor cortex (vSMC) and superior temporal gyrus (STG) and (\romannumeral2) self-supervision through discrete codex-guided mask modeling, significantly contribute to this performance. Overall, our approach -- inspired by neuroscience findings and capitalizing on region-level representations from specific brain regions -- is suitable for invasive brain modeling and represents a promising neuro-inspired AI approach in brain-computer interfaces. Code and dataset are available at https://github.com/liulab-repository/Du-IN.

In this paper, we propose a Mask-based Recipe Embedding GAN (MRE-GAN), which enables us to generate a realistic food image based on a given mask image containing single or multiple food regions with cross-modal recipe embeddings for each food region. Thus, we can change meal shapes by modifying mask images, while by editing recipe text, we can change meal appearance. Our experimental findings confirmed that the proposed method could generate higher quality food images than the baselines, and we could change meal shapes and appearances by editing mask images and recipe texts as we liked.