The Explanatory Effect of a Label: Its Influence on a Category Persists Even If We Forget the Label

In this study we replicated the explanatory effect of a label which had been found by Giffin et al. (2017). In their experiments, they used vignettes describing an odd behavior of a person based on culturally specific disorders that were unfamiliar to respondents. It turned out that explanations which explain an odd behavior through a person’s tendency to behave that way (circulus vitiosus) seemed more persuasive if the disorder was given a label that was used in the explanation. We replicated these results in Experiment 1, and in a follow-up Experiment 2 we examined the familiarity with category information and the evaluation of that category over time (the delay lasted one week). We realized that the label effect persists even when people make judgments based on their recollections about a category. Furthermore, according to a content analysis of the recollections, participants in the label condition remembered more information from the vignettes but tended to forget an artificial label; however, they used other words from the disorder domain instead (like “disease” or “kleptomania”). This allowed us to suggest a new interpretation of this effect: we suppose that in the Giffin et al. (2017) experiments the label did not bring any new features to a category itself, but pointed to a relevant domain instead, so the effect appeared from the activation of areas of knowledge in semantic memory and the application of relevant schema for learning a new phenomenon.

Attention enhances category representations across the brain with strengthened residual correlations to ventral temporal cortex

How does attention enhance neural representations of goal-relevant stimuli while suppressing representations of ignored stimuli across regions of the brain? While prior studies have shown that attention enhances visual responses, we lack a cohesive understanding of how selective attention modulates visual representations across the brain. Here, we used functional magnetic resonance imaging (fMRI) while participants performed a selective attention task on superimposed stimuli from multiple categories and used a data-driven approach to test how attention affects both decodability of category information and residual correlations (after regressing out stimulus-driven variance) with category-selective regions of ventral temporal cortex (VTC). Our data reveal three main findings. First, when two objects are simultaneously viewed, the category of the attended object can be decoded more readily than the category of the ignored object, with the greatest attentional enhancements observed in occipital and temporal lobes. Second, after accounting for the response to the stimulus, the correlation in the residual brain activity between a cortical region and a category-selective region of VTC was elevated when that region's preferred category was attended vs. ignored, and more so in the right occipital, parietal, and frontal cortices. Third, we found that the stronger the residual correlations between a given region of cortex and VTC, the better visual category information could be decoded from that region. These findings suggest that heightened residual correlations by selective attention may reflect the sharing of information between sensory regions and higher-order cortical regions to provide attentional enhancement of goal-relevant information.

The Effect of Spatial Frequency on the Visual Category Representation in the Macaque Inferior Temporal Cortex

ABSTRACTTo expand our knowledge about the object recognition, it is critical to understand the role of spatial frequency (SF) in an object representation that occurs in the inferior temporal (IT) cortex at the final stage of processing the visual information across the ventral visual pathway. Object categories are being recognized hierarchically in at least three levels of abstraction: superordinate (e.g., animal), mid-level (e.g., human face), and subordinate (e.g., face identity). Psychophysical studies have shown rapid access to mid-level category information and low SF (LSF) contents. Although the hierarchical representation of categories has been shown to exist inside the IT cortex, the impact of SF on the multi-level category processing is poorly understood. To gain a deeper understanding of the neural basis of the interaction between SF and category representations at multiple levels, we examined the neural responses within the IT cortex of macaque monkeys viewing several SF-filtered objects. Each stimulus could be either intact or bandpass filtered into either the LSF (coarse shape information) or high SF (HSF) (fine shape information) bands. We found that in both High- and Low-SF contents, the advantage of mid-level representation has not been violated. This evidence suggests that mid-level category boundary maps are strongly represented in the IT cortex and remain unaffected with respect to any changes in the frequency content of stimuli. Our observations indicate the necessity of the HSF content for the superordinate category representation inside the IT cortex. In addition, our findings reveal that the representation of global category information is more dependent on the HSF than the LSF content. Furthermore, the lack of subordinate representation in both LSF and HSF filtered stimuli compared to the intact stimuli provide strong evidence that all SF contents are necessary for fine category visual processing.

Temporal Variabilities Provide Additional Category-Related Information in Object Category Decoding: A Systematic Comparison of Informative EEG Features

How does the human brain encode visual object categories? Our understanding of this has advanced substantially with the development of multivariate decoding analyses. However, conventional electroencephalography (EEG) decoding predominantly uses the mean neural activation within the analysis window to extract category information. Such temporal averaging overlooks the within-trial neural variability that is suggested to provide an additional channel for the encoding of information about the complexity and uncertainty of the sensory input. The richness of temporal variabilities, however, has not been systematically compared with the conventional mean activity. Here we compare the information content of 31 variability-sensitive features against the mean of activity, using three independent highly varied data sets. In whole-trial decoding, the classical event-related potential (ERP) components of P2a and P2b provided information comparable to those provided by original magnitude data (OMD) and wavelet coefficients (WC), the two most informative variability-sensitive features. In time-resolved decoding, the OMD and WC outperformed all the other features (including the mean), which were sensitive to limited and specific aspects of temporal variabilities, such as their phase or frequency. The information was more pronounced in the theta frequency band, previously suggested to support feedforward visual processing. We concluded that the brain might encode the information in multiple aspects of neural variabilities simultaneously such as phase, amplitude, and frequency rather than mean per se. In our active categorization data set, we found that more effective decoding of the neural codes corresponds to better prediction of behavioral performance. Therefore, the incorporation of temporal variabilities in time-resolved decoding can provide additional category information and improved prediction of behavior.

YOLOMask, an Instance Segmentation Algorithm Based on Complementary Fusion Network

Object detection and segmentation can improve the accuracy of image recognition, but traditional methods can only extract the shallow information of the target, so the performance of algorithms is subject to many limitations. With the development of neural network technology, semantic segmentation algorithms based on deep learning can obtain the category information of each pixel. However, the algorithm cannot effectively distinguish each object of the same category, so YOLOMask, an instance segmentation algorithm based on complementary fusion network, is proposed in this paper. Experimental results on public data sets COCO2017 show that the proposed fusion network can accurately obtain the category and location information of each instance and has good real-time performance.

Towards a Reliable Evaluation of Local Interpretation Methods

The growing use of deep neural networks in critical applications is making interpretability urgently to be solved. Local interpretation methods are the most prevalent and accepted approach for understanding and interpreting deep neural networks. How to effectively evaluate the local interpretation methods is challenging. To address this question, a unified evaluation framework is proposed, which assesses local interpretation methods from three dimensions: accuracy, persuasibility and class discriminativeness. Specifically, in order to assess correctness, we designed an interactive user feature annotation tool to provide ground truth for local interpretation methods. To verify the usefulness of the interpretation method, we iteratively display part of the interpretation results, and then ask users whether they agree with the category information. At the same time, we designed and built a set of evaluation data sets with a rich hierarchical structure. Surprisingly, one finding is that the existing visual interpretation methods cannot satisfy all evaluation dimensions at the same time, and each has its own shortcomings.

Do Preschoolers Align Their Preferences With Those of a Powerful Individual?

Very early on, children understand the hierarchical dimension of the social environment and use a variety of cues to guess who has more power in an interaction. A crucial aspect of power perception lies in the evaluation of high-power and low-power individuals. The current study examined the evaluation of power by preschoolers through social influence. Past research has shown that preschoolers take social category information into account when expressing their preferences. In particular, they tend align their preferences with those of same-gender and same-age individuals. In the current study, 4- and 5-year-old children were presented with a power interaction between two children through body postures and were asked whether they would prefer the same items as those preferred by the high-power child or those preferred by the low-power child. Overall, the participants did not choose the items preferred by the high-power child significantly more often than those preferred by the low-power child. However, unexpected gender effects were found and indicated that the power asymmetry influenced more male than female participants. Indeed, when they saw a power interaction between two boys (Experiments 1 and 2), male participants aligned their choices with those of the high-power boy more than with those of the low-power boy. However, when male participants saw an interaction between two girls (Experiment 3), an opposite pattern was observed: they aligned their choices with those of the low-power girl more than with those of the high-power girl. In contrast, in the three experiments, there were approximately as many girls who aligned their preferences with those of the high-power child as there were girls who aligned their preferences with those of the low-power child. The current study reveals the importance of taking gender into account, both at the level of participants and stimuli, in the evaluation of power by preschoolers.