Search for: All records

Modeling cross-lingual speech emotion recognition (SER) has become more prevalent because of its diverse applications. Existing studies have mostly focused on technical approaches that adapt the feature, domain, or label across languages, without considering in detail the similarities be- tween the languages. This study focuses on domain adaptation in cross-lingual scenarios using phonetic constraints. This work is framed in a twofold manner. First, we analyze emotion-specific phonetic commonality across languages by identifying common vowels that are useful for SER modeling. Second, we leverage these common vowels as an anchoring mechanism to facilitate cross-lingual SER. We consider American English and Taiwanese Mandarin as a case study to demonstrate the potential of our approach. This work uses two in-the-wild natural emotional speech corpora: MSP-Podcast (American English), and BIIC-Podcast (Taiwanese Mandarin). The proposed unsupervised cross-lingual SER model using these phonetical anchors outperforms the baselines with a 58.64% of unweighted average recall (UAR). 

Advancing speech emotion recognition (SER) de- pends highly on the source used to train the model, i.e., the emotional speech corpora. By permuting different design parameters, researchers have released versions of corpora that attempt to provide a better-quality source for training SER. In this work, we focus on studying communication modes of collection. In particular, we analyze the patterns of emotional speech collected during interpersonal conversations or monologues. While it is well known that conversation provides a better protocol for eliciting authentic emotion expressions, there is a lack of systematic analyses to determine whether conversational speech provide a “better-quality” source. Specifically, we examine this research question from three perspectives: perceptual differences, acoustic variability and SER model learning. Our analyses on the MSP- Podcast corpus show that: 1) rater’s consistency for conversation recordings is higher when evaluating categorical emotions, 2) the perceptions and acoustic patterns observed on conversations have properties that are better aligned with expected trends discussed in emotion literature, and 3) a more robust SER model can be trained from conversational data. This work brings initial evidences stating that samples of conversations may provide a better-quality source than samples from monologues for building a SER model. 

Diabetes is one of the most pressing healthcare challenges facing society. Dysfunctional insulin signaling causes diabetes, leading to blood glucose instability and many associated complications. While the administration of exogenous insulin is then essential for achieving glucose control, issues with dosing accuracy and timing remain. Hydrogel‐based drug delivery systems have been broadly explored for controlled protein release, including for applications in long‐lasting and oral insulin delivery. More recently, efforts have focused on injectable hydrogels with glucose‐directed controlled release of insulin and glucagon, aiming for more autonomous and biomimetic approaches to blood glucose control. These materials typically use protein‐based sensing mechanisms or glucose binding by synthetic aryl boronates for glucose‐directed release. Despite advancements in this area, there remains a need for more precise timing of therapeutic availability to afford healthy blood glucose homeostasis, providing an opportunity for further research and innovation. This review summarizes the current state of hydrogel‐based delivery of insulin and glucagon, with insights into the potential benefits, future directions, and challenges that must be overcome to achieve clinical impact.

 

Hydrogels prepared from supramolecular cross-linking motifs are appealing for use as biomaterials and drug delivery technologies. The inclusion of macromolecules (e.g., protein therapeutics) in these materials is relevant to many of their intended uses. However, the impact of dynamic network cross-linking on macromolecule diffusion must be better understood. Here, hydrogel networks with identical topology but disparate cross-link dynamics are explored. These materials are prepared from cross-linking with host–guest complexes of the cucurbit[7]uril (CB[7]) macrocycle and two guests of different affinity. Rheology confirms differences in bulk material dynamics arising from differences in cross-link thermodynamics. Fluorescence recovery after photobleaching (FRAP) provides insight into macromolecule diffusion as a function of probe molecular weight and hydrogel network dynamics. Together, both rheology and FRAP enable the estimation of the mean network mesh size, which is then related to the solute hydrodynamic diameters to further understand macromolecule diffusion. Interestingly, the thermodynamics of host–guest cross-linking are correlated with a marked deviation from classical diffusion behavior for higher molecular weight probes, yielding solute aggregation in high-affinity networks. These studies offer insights into fundamental macromolecular transport phenomena as they relate to the association dynamics of supramolecular networks. Translation of these materials from in vitro to in vivo is also assessed by bulk release of an encapsulated macromolecule. Contradictory in vitro to in vivo results with inverse relationships in release between the two hydrogels underscores the caution demanded when translating supramolecular biomaterials into application. 

The transient self‐assembly of molecules under the direction of a consumable fuel source is fundamental to biological processes such as cellular organization and motility. Such biomolecular assemblies exist in an out‐of‐equilibrium state, requiring continuous consumption of high energy molecules. At the same time, the creation of bioinspired supramolecular hydrogels has traditionally focused on associations occurring at the thermodynamic equilibrium state. Here, hydrogels are prepared from cucurbit[7]uril host–guest supramolecular interactions through transient physical crosslinking driven by the consumption of a reactive chemical fuel. Upon action from this fuel, the affinity and dynamics of CB[7]–guest recognition are altered. In this way, the lifetime of transient hydrogel formation and the dynamic modulus obtained are governed by fuel consumption, rather than being directed by equilibrium complex formation.