Search for: All records

Multimodal fusion addresses the problem of analyzing spoken words in the multimodal context, including visual expressions and prosodic cues. Even when multimodal models lead to performance improvements, it is often unclear whether bimodal and trimodal interactions are learned or whether modalities are processed independently of each other. We propose Multimodal Residual Optimization (MRO) to separate unimodal, bimodal, and trimodal interactions in a multimodal model. This improves interpretability as the multimodal interaction can be quantified. Inspired by Occam’s razor, the main intuition of MRO is that (simpler) unimodal contributions should be learned before learning (more complex) bimodal and trimodal interactions. For example, bimodal predictions should learn to correct the mistakes (residuals) of unimodal predictions, thereby letting the bimodal predictions focus on the remaining bimodal interactions. Empirically, we observe that MRO successfully separates unimodal, bimodal, and trimodal interactions while not degrading predictive performance. We complement our empirical results with a human perception study and observe that MRO learns multimodal interactions that align with human judgments. 

RationaleFemtoamp and picoamp electrospray ionization (ESI) characteristics of a nonpolar solvent were explored. The direct ESI mass spectrometry analysis of chloroform extract solution enabled rapid analysis of perfluorinated sulfonic acid analytes in drinking water. MethodsNeat chloroform solvent and extracts were directly used in a typical wire‐in ESI setup using micrometer emitter tips. Ionization currents were measured with femtoamp sensitivity while ramping the spray voltage from 0 to −5000 V. Methanol was used as a comparison to illustrate the characteristics of electrospraying chloroform. The effects of spray voltage and inlet temperature were studied. A liquid–liquid extraction workflow was developed to analyze perfluorooctanoate sulfonate (PFOS) in drinking water using an ion‐trap mass spectrometer. ResultsThe ionization onset of chloroform solution was 41 ± 17 fA at 300 V. The ionization current gradually increased with voltage while remaining below 100 pA when using voltages up to −5000 V. The ion signal of PFOS was significantly enhanced to improve the limit of detection (LoD) to 25 ppt in chloroform. Coupled with a liquid–liquid extraction workflow, LoD of 0.38–5.1 ppt and a quantitation range of 5–400 ppt were achieved for perfluorinated sulfonic compounds in 1‐ml water samples. ConclusionsFemtoamp and picoamp modes expand the solvent compatibility range of ESI and can enable quantitative analysis in parts per trillion (ppt) concentrations. 

This paper studies the hypothesis that not all modalities are always needed to predict affective states. We explore this hypothesis in the context of recognizing three affective states that have shown a relation to a future onset of depression: positive, aggressive, and dysphoric. In particular, we investigate three important modali- ties for face-to-face conversations: vision, language, and acoustic modality. We first perform a human study to better understand which subset of modalities people find informative, when recog- nizing three affective states. As a second contribution, we explore how these human annotations can guide automatic affect recog- nition systems to be more interpretable while not degrading their predictive performance. Our studies show that humans can reliably annotate modality informativeness. Further, we observe that guided models significantly improve interpretability, i.e., they attend to modalities similarly to how humans rate the modality informative- ness, while at the same time showing a slight increase in predictive performance.