Emotion words help children to learn about emotions, yet the mechanism behind this remains unclear. These words may be helpful to children because they are novel, but this emotion learning may also be driven by familiarity. To address this, we conducted a pre-test post-test examination of emotion learning (matching faces to scenarios) with 3-year-olds (N=72). Between pre- and post-test, children either heard explicit emotion labels (“she feels annoyed”), novel labels (“she feels wuggy”), or irrelevant information (“she sits down”). Results revealed that children’s change score (post-test – pre-test) significantly differed by condition (F(2,69)=4.51, p=.014). The most positive change occurred in the explicit label condition (M=0.96, SD=1.57) followed by novel labels (M=0.08, SD=1.69) and irrelevant information (M=-0.38, SD=1.41). Children learned when provided with explicit emotion labels, but not novel labels. This suggests that familiarity (e.g., from previous experience), and not novelty, may best explain how emotion words influence children’s learning about emotions.

How Do Emotion Labels Help Emotion Learning? The Role of Novelty Versus Familiarity

Children are able to associate a word with its referent even after a single exposure (i.e., fast-mapping). This ability plays a central role in vocabulary development. Often, words are learned in context and not in isolation. However, most word-learning studies have not considered visual context cues. In the present eye-tracking study we investigated the effects of contextual information (constant versus variable context) on word-referent association (exposure phase) and learning (test phase) during fast-mapping in school-age children. Also, we examined how vocabulary and attention skills modulate this process. Our results show that constant visual context facilitated novel word-reference association (exposure phase) in both age groups while word learning (test phase) was modulated by the visual context only in the younger group. Similarly, visual attention skills ease word learning in the most demanding condition (variable condition) in this group only. Finally, vocabulary skills predicted word learning in the older group.

Visual Context and Word Learning in School-Age Children: An Eye-tracking Study

This study attempts to gain novel insights into the mechanisms underlying the bilingual advantage in cognitive control and investigates the effect of code-switching experience in daily life, rather than considering only language proficiency, on these underlying mechanisms. We compared the performances of early and late Japanese-English bilinguals on the AX-Continuous Performance Task. We found early and late bilinguals show different cognitive control mechanisms to cope with interference: with the late bilinguals, the more frequently they switched languages, the better they performed in the proactive control tasks. Conversely, with the early bilinguals who are generally more proficient, the frequency of code-switching did not modulate their dual mechanism of cognitive control any better, showing little correlation with the ability of proactive control. These findings verify the role of switching frequency in the development of cognitive control and suggest a better understanding of the nature of mechanisms regulating early and late bilinguals’ brains.

The effect of code-switching experience on the dual mechanism of cognitive control: A comparison between early and late Japanese-English bilinguals

Past studies have revealed that language experience impacts children’s vocabulary development. For example, bilingual children tend to have smaller vocabularies than monolingual children in each of their languages. However, it is unclear whether routine exposure to multiple accents also affects children’s vocabulary growth. Here, using standardized vocabulary assessments, we compared the reported vocabulary sizes of 11- to 34-month-old monolingual and bilingual children (N = 2881) who had various degrees of accent exposure. Our results show that routine exposure to multiple accents, regardless of accent type, does not negatively impact vocabulary development. Our findings suggest that children are well-equipped to handle language variation in their input.

Effects of Accent Exposure on Monolingual and Bilingual Children’s Early Vocabulary Development

Active learning techniques enhance understanding and engagement in the learner, but these techniques can be difficult to teach to instructors. The pedagogy of active learning typically either describes abstract concepts or it relies heavily on case studies of what other instructors have created. Because it can be difficult for interested instructors to put theory into practice without a concrete structure, active learning techniques can be hard to implement in the classroom. Behaviorism, cognitivism, constructivism, and connectivism can be used to describe ideas about how learning is occurring, but they do not accurately categorize commonly used teaching activities, which instead often contain elements of multiple learning theories. I propose a novel functional categorization of active learning techniques that is both brain-based and student-centered, placing emphasis on the knowledge journey. This work describes these seven evidence-based “knowledge catalysts” as a new user-friendly framework for both understanding, using, and teaching active learning techniques.

Knowledge Catalysts: A Structure to Promote Incorporation of Active Learning Techniques

Deep reinforcement learning (DRL) is capable of training agents that exceed human-levels of performance in multi-agent tasks. However, the behaviors exhibited by these agents are not guaranteed to be human-like or human-compatible. This poses a problem if the goal is to design agents capable of collaborating with humans or augmenting human actions in cooperative or team-based tasks. Indeed, recommender systems designed to augment human decision making need to not only recommend actions that align with the task goal, but which also maintain coordinative behaviors between agents. The current study simultaneously explored skill learning performance of human learners when working alongside different artificial agents (AAs) during a collaborative problem-solving task, as well as evaluated the effectiveness of the same AAs as action decision recommender systems to aid learning. The action decisions of the AAs were either modelled by a heuristic model based on human performance or by a deep neural network trained by reinforcement learning in self-play. In addition to evaluating skill learning performance, the current study also tested the congruence between the decisions of the AAs with actual decisions made by humans. Results demonstrate that the performance of humans was significantly worse when working alongside the DRL AA compared to the heuristic AA. Additionally, the action decisions participants made also showed less allignment with the recommendations made by the DRL AA.

Action decision congruence between human and deep reinforcement learning agents during a coordinated action task

People are known for their ability to learn probabilistic rewarding rules of the environment. However, the rules tested in most studies consisted of one feature dimension or a linear combination of multiple features. It is unclear how well people can learn non-linear combinatory rules where each single dimension is noninformative. In an adapted matching-pennies game, we asked participants to predict whether their opponent would choose head or tail next trial. Though the feature space for possible rules was small, including only two binary dimensions (i.e., the two players’ current choices), participants had striking failures in learning the probabilistic XOR rules (e.g., the opponent would have a probability of 0.8 to choose head if the two players had matching pennies and tail for unmatching pennies). The failure persisted in a series of experiments, even when the feature space was made explicit, after 320 trials of learning, and for different cover stories.

People fail to learn probabilistic XOR rules even in a small feature space

In recent years, online lending platforms have been becoming attractive for micro-financing and popular in financial industries. However, such online lending platforms face a high risk of failure due to the lack of expertise on borrowers' creditworthness. Thus, risk forecasting is important to avoid economic loss. Detecting loan fraud users in advance is at the heart of risk forecasting. The purpose of fraud user (borrower) detection is to predict whether one user will fail to make required payments in the future. Detecting fraud users depend on historical loan records. However, a large proportion of users lack such information, especially for new users. In this paper, we attempt to detect loan fraud users from cross domain heterogeneous data views, including user attributes, installed app lists, app installation behaviors, and app-in logs, which compensate for the lack of historical loan records. However, it is difficult to effectively fuse the multiple heterogeneous data views. Moreover, some samples miss one or even more data views, increasing the difficulty in fusion. To address the challenges, we propose a novel end-to-end deep multiview learning approach, which encodes heterogeneous data views into homogeneous ones, generates the missing views based on the learned relationship among all the views, and then fuses all the views together to a comprehensive view for identifying fraud users. Our model is evaluated on a real-world large-scale dataset consisting of 401,978 loan records of 228,117 users from January 1, 2019, to September 30, 2019, achieving the state-of-the-art performance. 

Loan Fraud Users Detection in Online Lending Leveraging Multiple Data Views

In recent years, multi-view multi-label learning has aroused extensive research enthusiasm. However, multi-view multi-label data in the real world is commonly incomplete due to the uncertain factors of data collection and manual annotation, which means that not only multi-view features are often missing, and label completeness is also difficult to be satisfied. To deal with the double incomplete multi-view multi-label classification problem, we propose a deep instance-level contrastive network, namely DICNet. Different from conventional methods, our DICNet focuses on leveraging deep neural network to exploit the high-level semantic representations of samples rather than shallow-level features. First, we utilize the stacked autoencoders to build an end-to-end multi-view feature extraction framework to learn the view-specific representations of samples. Furthermore, in order to improve the consensus representation ability, we introduce an incomplete instance-level contrastive learning scheme to guide the encoders to better extract the consensus information of multiple views and use a multi-view weighted fusion module to enhance the discrimination of semantic features. Overall, our DICNet is adept in capturing consistent discriminative representations of multi-view multi-label data and avoiding the negative effects of missing views and missing labels. Extensive experiments performed on five datasets validate that our method outperforms other state-of-the-art methods.

DICNet: Deep Instance-Level Contrastive Network for Double Incomplete Multi-View Multi-Label Classification

Deep Reinforcement Learning (DRL) has the potential to be used for synthesizing feedback controllers (agents) for various complex systems with unknown dynamics. These systems are expected to satisfy diverse safety and liveness properties best captured using temporal logic. In RL, the reward function plays a crucial role in specifying the desired behaviour of these agents. However, the problem of designing the reward function for an RL agent to satisfy complex temporal logic specifications has received limited attention in the literature. To address this, we provide a systematic way of generating rewards in real-time by using the quantitative semantics of Signal Temporal Logic (STL), a widely used temporal logic to specify the behaviour of cyber-physical systems. We propose a new quantitative semantics for STL having several desirable properties, making it suitable for reward generation. We evaluate our STL-based reinforcement learning mechanism on several complex continuous control benchmarks and compare our STL semantics with those available in the literature in terms of their efficacy in synthesizing the controller agent. Experimental results establish our new semantics to be the most suitable for synthesizing feedback controllers for complex continuous dynamical systems through reinforcement learning.

STL-Based Synthesis of Feedback Controllers Using Reinforcement Learning

One approach to guaranteeing safety in Reinforcement Learning is through cost constraints that are dependent on the policy. Recent works in constrained RL have developed methods that ensure  constraints are enforced even at learning time while maximizing the overall value of the policy. Unfortunately, as demonstrated in our experimental results, such approaches do not perform well on complex multi-level tasks, with longer episode lengths or sparse rewards. To that end, we propose a scalable hierarchical approach for constrained RL problems that employs backward cost value functions in the context of task hierarchy and a novel intrinsic reward function in lower levels of the hierarchy to enable cost constraint enforcement. One of our key contributions is in proving that backward value functions are theoretically viable even when there are multiple levels of decision making. We also show that our new approach, referred to as Hierarchically Limited consTraint Enforcement (HiLiTE) significantly improves on state of the art Constrained RL approaches for many  benchmark problems from literature. We further demonstrate that this performance (on value and constraint enforcement) clearly outperforms existing best approaches for constrained RL and hierarchical RL. 

Constrained Reinforcement Learning in Hard Exploration Problems

Standard deep reinforcement learning (DRL) aims to maximize expected reward, considering collected experiences equally in formulating a policy. This differs from human decision-making, where gains and losses are valued differently and outlying outcomes are given increased consideration.  It also fails to capitalize on opportunities to improve safety and/or performance through the incorporation of distributional context. Several approaches to distributional DRL have been investigated, with one popular strategy being to evaluate the projected distribution of returns for possible actions.  We propose a more direct approach whereby risk-sensitive objectives, specified in terms of the cumulative distribution function (CDF) of the distribution of full-episode rewards, are optimized. This approach allows for outcomes to be weighed based on relative quality, can be used for both continuous and discrete action spaces, and may naturally be applied in both constrained and unconstrained settings.  We show how to compute an asymptotically consistent estimate of the policy gradient for a broad class of risk-sensitive objectives via sampling, subsequently incorporating variance reduction and regularization measures to facilitate effective on-policy learning.  We then demonstrate that the use of moderately "pessimistic" risk profiles, which emphasize scenarios where the agent performs poorly, leads to enhanced exploration and a continual focus on addressing deficiencies.  We test the approach using different risk profiles in six OpenAI Safety Gym environments, comparing to state of the art on-policy methods.  Without cost constraints, we find that pessimistic risk profiles can be used to reduce cost while improving total reward accumulation.  With cost constraints, they are seen to provide higher positive rewards than risk-neutral approaches at the prescribed allowable cost.

Underline digital video library

How Do Emotion Labels Help Emotion Learning? The Role of Novelty Versus Familiarity

Visual Context and Word Learning in School-Age Children: An Eye-tracking Study

The effect of code-switching experience on the dual mechanism of cognitive control: A comparison between early and late Japanese-English bilinguals

Effects of Accent Exposure on Monolingual and Bilingual Children’s Early Vocabulary Development

Knowledge Catalysts: A Structure to Promote Incorporation of Active Learning Techniques

Action decision congruence between human and deep reinforcement learning agents during a coordinated action task

People fail to learn probabilistic XOR rules even in a small feature space

Loan Fraud Users Detection in Online Lending Leveraging Multiple Data Views

DICNet: Deep Instance-Level Contrastive Network for Double Incomplete Multi-View Multi-Label Classification

STL-Based Synthesis of Feedback Controllers Using Reinforcement Learning

Constrained Reinforcement Learning in Hard Exploration Problems

A Risk-Sensitive Approach to Policy Optimization

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES