Automatic radiology report generation has attracted enormous research interest due to its practical value in reducing the workload of radiologists. However, simultaneously establishing global correspondences between the image (e.g., Chest X-ray) and its related report and local alignments between image patches and keywords remains challenging. To this end, we propose an Unify, Align and then Refine (UAR) approach to learn multi-level cross-modal alignments and introduce three novel modules: Latent Space Unifier (LSU), Cross-modal Representation Aligner (CRA) and Text-to-Image Refiner (TIR). Specifically, LSU unifies multimodal data into discrete tokens, making it flexible to learn common knowledge among modalities with a shared network. The modality-agnostic CRA learns discriminative features via a set of orthonormal basis and a dual-gate mechanism first and then globally aligns visual and textual representations under a triplet contrastive loss. TIR boosts token-level local alignment via calibrating text-to-image attention with a learnable mask. Additionally, we design a two-stage training procedure to make UAR gradually grasp cross-modal alignments at different levels, which imitates radiologists’ workflow: writing sentence by sentence first and then checking word by word. Extensive experiments and analyses on IU-Xray and MIMIC-CXR benchmark datasets demonstrate the superiority of our UAR against varied state-of-the-art methods.

The recent video grounding works attempt to introduce vanilla contrastive learning into video grounding. However, we claim that this naive solution is suboptimal. Contrastive learning requires two key properties: (1) alignment of features of similar samples, and (2) uniformity of the induced distribution of the normalized features on the hypersphere. Due to two annoying issues in video grounding: (1) the coexistence of some visual entities in both ground truth and other moments, i.e. semantic overlapping; (2) only a few moments in the video are annotated, i.e. sparse annotation dilemma, vanilla contrastive learning is unable to model the correlations between temporally distant moments and learned inconsistent video representations. Both characteristics lead to vanilla contrastive learning being unsuitable for video grounding. In this paper, we introduce Geodesic and Game Localization (G2L), a semantically aligned and uniform video grounding framework via geodesic and game theory. We quantify the correlations among moments leveraging the geodesic distance that guides the model to learn the correct cross-modal representations. Furthermore, from the novel perspective of game theory, we propose semantic Shapley interaction based on geodesic distance sampling to learn fine-grained semantic alignment in similar moments. Experiments on three benchmarks demonstrate the effectiveness of our method.

The effects of sintering temperature and reoxidation conduction on the positive temperature coefficient of resistivity (PTCR) effect of chip type Ba_1.005-xSm_xTiO₃ (BSMT) ceramics were investigated. The samples were reoxidized at 600°-800°C for 0-3 h in air after being sintered at 1200°-1280°C for 30 min in a reducing atmosphere. The results indicated that the reduced BSMT ceramics reoxidized at 800°C exhibited a pronounced resistivity jump. Moreover, the room temperature resistivity and resistivity jump of samples decreased with increasing sintering temperature of 1200°-1280°C. Furthermore, BSMT ceramics with 0.6 mol% Sm³⁺ reoxidized at 750°C for 3 h after sintering at 1200°C for 30 min exhibited a significant resistivity jump of 3.2 orders of magnitude, along with a relatively low room temperature resistivity of 661.4 Ω.cm. In addition, the activation energy of the samples obtained at different sintering temperatures was also investigated.