Evaluating pedestrian interaction preferences with a game theoretic autonomous vehicle in virtual reality

Localisation and navigation of autonomous vehicles (AVs) in static environments are now solved problems, but how to control their interactions with other road users in mixed traffic environments, especially with pedestrians, remains an open question. Recent work has begun to apply game theory to model and control AV-pedestrian interactions as they compete for space on the road whilst trying to avoid collisions. But this game theory model has been developed only in unrealistic lab environments. To improve their realism, this study empirically examines pedestrian behaviour during road crossing in the presence of approaching autonomous vehicles in more realistic virtual reality (VR) environments. The autonomous vehicles are controlled using game theory, and this study seeks to find the best parameters for these controls to produce comfortable interactions for the pedestrians. In a first experiment, participants’ trajectories reveal a more cautious crossing behaviour in VR than in previous laboratory experiments. In two further experiments, a gradient descent approach is used to investigate participants’ preference for AV driving style. The results show that the majority of participants were not expecting the AV to stop in some scenarios, and there was no change in their crossing behaviour in two environments and with different car models suggestive of car and last-mile style vehicles. These results provide some initial estimates for game theoretic parameters needed by future AVs in their pedestrian interactions and more generally show how such parameters can be inferred from virtual reality experiments.     

EARSHOT: A Minimal Neural Network Model of Incremental Human Speech Recognition

Despite the lack of invariance problem (the many-to-many mapping between acoustics and percepts), human listeners experience phonetic constancy and typically perceive what a speaker intends. Most models of human speech recognition (HSR) have side-stepped this problem, working with abstract, idealized inputs and deferring the challenge of working with real speech. In contrast, carefully engineered deep learning networks allow robust, real-world automatic speech recognition (ASR). However, the complexities of deep learning architectures and training regimens make it difficult to use them to provide direct insights into mechanisms that may support HSR. In this brief article, we report preliminary results from a two-layer network that borrows one element from ASR, long short-term memory nodes, which provide dynamic memory for a range of temporal spans. This allows the model to learn to map real speech from multiple talkers to semantic targets with high accuracy, with human-like timecourse of lexical access and phonological competition. Internal representations emerge that resemble phonetically organized responses in human superior temporal gyrus, suggesting that the model develops a distributed phonological code despite no explicit training on phonetic or phonemic targets. The ability to work with real speech is a major advance for cognitive models of HSR.     

Human factors evaluation of an in-vehicle active traffic and demand management (ATDM) system

This research study focused on the evaluation of an emulated in-vehicle Active Traffic and Demand Management (ATDM) system on Interstate 66 in Northern Virginia. Traditional ATDM systems rely on infrastructure-mounted variable message signs (VMS) to provide information (speed limits, lane availability, etc.) to the traveling public. By providing information about dynamic roadway conditions on an in-vehicle device, the ATDM may improve driving safety and performance by allowing drivers to remain consistently aware of forthcoming traffic conditions and roadway requirements; even when external signage is inaccessible. This study represents an initial investigation of an emulated in-vehicle ATDM to provide developers with design guidance and ensure that unintended consequences, such as distraction, do not undermine the potential benefits.Twenty younger and 20 older participants, accompanied by a member of the research team, experienced the following ATDM features on an in-vehicle device (IVD) mounted to the dashboard: (1) dynamic speed limits, (2) dynamic lane use/shoulder control, (3) High Occupancy Vehicle (HOV) restrictions, and (4) roadway information from variable message signs. The IVD was equipped with auditory and visual alerts notifying the driver when relevant visual information was updated. Research questions addressed distraction and driver behavior associated with use of the system. Qualitative and quantitative participant data was acquired from the instrumented vehicle, various questionnaires, and researcher observation.Several key findings were uncovered: (1) The IVD, as tested, did not warrant classification as a source of distraction according to the NHTSA guidelines; v2) There was a significant difference in eye-glance durations to the IVD when comparing the VMS alerts to both the speed limit and lane management alerts; and (3) The speed limit alert motivated participants to alter their speed (per survey results and participant speed data).     

Infants' understanding of object-directed action

When and in what ways do infants recognize humans as intentional actors? An important aspect of this larger question concerns when infants recognize specific human actions (e.g. a reach) as object-directed (i.e. as acting toward goal-objects). In two studies using a visual habituation technique, 12-month-old infants were tested to assess their recognition that an adult's reach is directed toward its target object. Infants in the experimental condition were habituated to a display in which an actor reached over a wall-like barrier with an arcing arm movement, to pick up a ball. After habituation infants saw two test displays, for which the barrier was removed. In the direct test event the actor reached directly for the ball, the arm tracing a visually new path, but the action consistent with attempting to reach for the object as directly as possible. In the indirect test event the actor traced the old path, reaching over in an arc, even though the wall was no longer present. This arm movement was identical to that in habituation but no longer displayed a reach going directly to its object. In a control condition infants saw the same movements but in a situation with no goal-object. In the experimental conditions, with a goal object present, infants looked longer at the indirect test event in comparison to the direct test event. In the control conditions infants looked equally at both indirect and direct test events. We conclude that sensitivity to human object-directed action is established by 12-month-olds and compare these results to recent findings by [Gergely, G., Nadasdy, Z., Csibra, G., & Biro S. (1995). Taking the intentional stance at 12 months of age. Cognition, 56, 165-193] and [Woodward, A. (1998). Infants selectively encode the goal object of an actor's reach. Cognition, 69, 1-34].     

近五十年来台湾地区士林哲学之研究与前瞻

根据祁尔松（Etienne Gilson）的说法,中世纪哲学,尤其是十三世纪的思想,是基督宗教哲学。本文按其在台湾发展时所袭用的名词,称之为士林哲学。士林哲学之传入中国,虽然早在利玛窦时期（1550—1610）已经开始,但真正成为学术思想之主流之一,是最近五十年来在台湾地区的发展。20世纪初士林哲学初传入台湾时,当时的思想环境,对基督宗教哲学并不十分友善,这种种情况,与中世纪的思想家,面对信仰与神学之挑战相似。当时的思想家,融合各种不同的思想派系,对柏拉图与亚理斯多德的观念,重新诠释与批判,批判与诠释是创新的基础。所以,中世纪可说是哲学的第二春,而中世纪哲学之精华,即全盛期的士林哲学。十四世纪唯名论之后,士林哲学逐步式微,于十九世纪末再兴。本文认为,士林哲学在台湾发展之过程,与其在中世纪之盛行,及十九世纪末之再兴,背景虽异但有不少相似之处,而这些相似之处,正是士林哲学之基本思想与立场。本文对台湾士林哲学之发展,从四方面探讨：（1）回到形上学之根以面对新的挑战;（2）形上学基本立场与概念之说明;（3）反思与批判;（4）未来发展之方向。本文所参考之资料,一是已发表之学术文献;其二是与学者们之正式交谈——如学术演讲、座谈会等,及非正式之谈话、访问等。     

An experimental study of the emergence of human communication systems

The emergence of human communication systems is typically investigated via 2 approaches with complementary strengths and weaknesses: naturalistic studies and computer simulations. This study was conducted with a method that combines these approaches. Pairs of participants played video games requiring communication. Members of a pair were physically separated but exchanged graphic signals through a medium that prevented the use of standard symbols (e.g., letters). Communication systems emerged and developed rapidly during the games, integrating the use of explicit signs with information implicitly available to players and silent behavior-coordinating procedures. The systems that emerged suggest 3 conclusions: (a) signs originate from different mappings; (b) sign systems develop parsimoniously; (c) sign forms are perceptually distinct, easy to produce, and tolerant to variations.     

Processing coordinated structures: incrementality and connectedness

We recorded participants' eye movements while they read sentences containing verb-phrase coordination. Results showed evidence of immediate processing disruption when a reflexive pronoun embedded in the conjoined verb phrase mismatched the sentence subject. We argue that this result is incompatible with models of human parsing that employ only bottom-up parsing procedures, even when flexible constituency is employed. Models need to incorporate a mechanism similar to the adjoining operation in Tree-Adjoining Grammar, in which one structure is inserted into another.     

How Should a Speech Recognizer Work?

Although researchers studying human speech recognition (HSR) and automatic speech recognition (ASR) share a common interest in how information processing systems (human or machine) recognize spoken language, there is little communication between the two disciplines. We suggest that this lack of communication follows largely from the fact that research in these related fields has focused on the mechanics of how speech can be recognized. In Marr's (1982) terms, emphasis has been on the algorithmic and implementational levels rather than on the computational level. In this article, we provide a computational-level analysis of the task of speech recognition, which reveals the close parallels between research concerned with HSR and ASR. We illustrate this relation by presenting a new computational model of human spoken-word recognition, built using techniques from the field of ASR that, in contrast to current existing models of HSR, recognizes words from real speech input.     

Size of the neocerebellar vermis is associated with cognition in healthy elderly men

Cerebellar volumes show small positive correlations with cognitive ability in young adults, but no studies have examined this relationship in older adults. Furthermore, no studies have examined relationships between sizes of subareas of the cerebellum and cognitive ability. We hypothesized that size of the two areas of the neocerebellar vermis would correlate with a battery of eight cognitive tests in 50 men aged 65-70. Size of the neocerebellar areas of the vermis correlated positively with several cognitive tests (r's=.29-.37, p<.05), whilst sizes of other parts of the vermis did not correlate with any cognitive tests. Total cerebellar volumes correlated significantly with a test of nonverbal reasoning (r=.030, p=.42). These findings suggest a specific association of neocerebellar vermis size with variations in cognitive ability in older adults.     

Differential memory-preserving effects of reminders at 6 months

Although reactivation and reinstatement reminders differ procedurally, differences in their memory-preserving effects have been described as artifactual. In three experiments, we examined this conclusion. One hundred and twelve 6-month-olds learned an operant task, forgot it, received a reactivation or reinstatement reminder to recover the inactive memory, and were tested after increasing delays until they forgot it again. In Experiments 1a and 1b, a single reactivation reminder extended infants' memory of an operant mobile task for 2 weeks after reminding, but a single reinstatement extended it for 4 weeks, when testing was discontinued. In Experiment 2, a single reinstatement extended 6-month-olds' memory of an operant train task for 19 weeks after reminding, when infants were almost 1 year old. After reactivation, infants remember this task for only 2 weeks. The finding that the memory-preserving effect of reinstatement is greater by an order of magnitude suggests that procedural differences between the two reminders have functional significance.