A major advance in understanding learning behavior stems from experiments showing that reward learning requires dopamine inputs to striatal neurons and arises from synaptic plasticity of cortico-striatal synapses. Numerous reinforcement learning models mimic this dopamine-dependent synaptic plasticity by using the reward prediction error, which resembles dopamine neuron firing, to learn the best action in response to a set of cues. Though these models can explain many facets of behavior, reproducing some types of goal-directed behavior, such as renewal and reversal, require additional model components. Here we present a reinforcement learning model, TD2Q, which better corresponds to the basal ganglia with two Q matrices, one representing direct pathway neurons (G) and another representing indirect pathway neurons (N). Unlike previous two-Q architectures, a novel and critical aspect of TD2Q is to update the G and N matrices utilizing the temporal difference reward prediction error. A best action is selected for N and G using a softmax with a reward-dependent adaptive exploration parameter, and then differences are resolved using a second selection step applied to the two action probabilities. The model is tested on a range of multi-step tasks including extinction, renewal, discrimination; switching reward probability learning; and sequence learning. Simulations show that TD2Q produces behaviors similar to rodents in choice and sequence learning tasks, and that use of the temporal difference reward prediction error is required to learn multi-step tasks. Blocking the update rule on the N matrix blocks discrimination learning, as observed experimentally. Performance in the sequence learning task is dramatically improved with two matrices. These results suggest that including additional aspects of basal ganglia physiology can improve the performance of reinforcement learning models, better reproduce animal behaviors, and provide insight as to the role of direct- and indirect-pathway striatal neurons.
The current study aimed to elucidate the contributions of the subcortical basal ganglia to human language by adopting the view that these structures engage in a basic neurocomputation that may account for its involvement across a wide range of linguistic phenomena. Specifically, we tested the hypothesis that basal ganglia reinforcement learning (RL) mechanisms may account for variability in semantic selection processes necessary for ambiguity resolution. To test this, we used a biased homograph lexical ambiguity priming task that allowed us to measure automatic processes for resolving ambiguity toward high‐frequency word meanings. Individual differences in task performance were then related to indices of basal ganglia RL, which were used to group subjects into three learning styles: (a) Choosers who learn by seeking high reward probability stimuli; (b) Avoiders, who learn by avoiding low reward probability stimuli; and (c) Balanced participants, whose learning reflects equal contributions of choose and avoid processes. The results suggest that balanced individuals had significantly lower access to subordinate, or low‐frequency, homograph word meanings. Choosers and Avoiders, on the other hand, had higher access to the subordinate word meaning even after a long delay between prime and target. Experimental findings were then tested using an ACT‐R computational model of RL that learns from both positive and negative feedback. Results from the computational model simulations confirm and extend the pattern of behavioral findings, providing an RL account of individual differences in lexical ambiguity resolution.
more » « less- NSF-PAR ID:
- 10133490
- Publisher / Repository:
- Wiley-Blackwell
- Date Published:
- Journal Name:
- Topics in Cognitive Science
- Volume:
- 12
- Issue:
- 1
- ISSN:
- 1756-8757
- Page Range / eLocation ID:
- p. 402-416
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
Cai, Ming Bo (Ed.)more » « less
-
more » « less
This exploratory study examined the simultaneous interactions and relative contributions of bottom-up social information (regional dialect, speaking style), top-down contextual information (semantic predictability), and the internal dynamics of the lexicon (neighborhood density, lexical frequency) to lexical access and word recognition. Cross-modal matching and intelligibility in noise tasks were conducted with a community sample of adults at a local science museum. Each task featured one condition in which keywords were presented in isolation and one condition in which they were presented within a multiword phrase. Lexical processing was slower and more accurate when keywords were presented in their phrasal context, and was both faster and more accurate for auditory stimuli produced in the local Midland dialect. In both tasks, interactions were observed among stimulus dialect, speaking style, semantic predictability, phonological neighborhood density, and lexical frequency. These interactions revealed that bottom-up social information and top-down contextual information contribute more to speech processing than the internal dynamics of the lexicon. Moreover, the relatively stronger bottom-up social effects were observed in both the isolated word and multiword phrase conditions, suggesting that social variation is central to speech processing, even in non-interactive laboratory tasks. At the same time, the specific interactions observed differed between the two experiments, reflecting task-specific demands related to processing time constraints and signal degradation.
-
Abstract To accurately categorize items, humans learn to selectively attend to the stimulus dimensions that are most relevant to the task. Models of category learning describe how attention changes across trials as labeled stimuli are progressively observed. The Adaptive Attention Representation Model (AARM), for example, provides an account in which categorization decisions are based on the perceptual similarity of a new stimulus to stored exemplars, and dimension-wise attention is updated on every trial in the direction of a feedback-based error gradient. As such, attention modulation as described by AARM requires interactions among processes of orienting, visual perception, memory retrieval, prediction error, and goal maintenance to facilitate learning. The current study explored the neural bases of attention mechanisms using quantitative predictions from AARM to analyze behavioral and fMRI data collected while participants learned novel categories. Generalized linear model analyses revealed patterns of BOLD activation in the parietal cortex (orienting), visual cortex (perception), medial temporal lobe (memory retrieval), basal ganglia (prediction error), and pFC (goal maintenance) that covaried with the magnitude of model-predicted attentional tuning. Results are consistent with AARM's specification of attention modulation as a dynamic property of distributed cognitive systems.more » « less
-
Many computational models of reasoning rely on explicit relation representations to account for human cognitive capacities such as analogical reasoning. Relational luring, a phenomenon observed in recognition memory, has been interpreted as evidence that explicit relation representations also impact episodic memory; however, this assumption has not been rigorously assessed by computational modeling. We implemented an established model of recognition memory, the Generalized Context Model (GCM), as a framework for simulating human performance on an old/new recognition task that elicits relational luring. Within this basic theoretical framework, we compared representations based on explicit relations, lexical semantics (i.e., individual word meanings), and a combination of the two. We compared the same alternative representations as predictors of accuracy in solving explicit verbal analogies. In accord with previous work, we found that explicit relation representations are necessary for modeling analogical reasoning. In contrast, preliminary simulations incorporating model parameters optimized to fit human data reproduce relational luring using any of the alternative representations, including one based on non-relational lexical semantics. Further work on model comparisons is needed to examine the contributions of lexical semantics and relations on the luring effect in recognition memory.more » « less
-
null (Ed.)The DeepLearningEpilepsyDetectionChallenge: design, implementation, andtestofanewcrowd-sourced AIchallengeecosystem Isabell Kiral*, Subhrajit Roy*, Todd Mummert*, Alan Braz*, Jason Tsay, Jianbin Tang, Umar Asif, Thomas Schaffter, Eren Mehmet, The IBM Epilepsy Consortium◊ , Joseph Picone, Iyad Obeid, Bruno De Assis Marques, Stefan Maetschke, Rania Khalaf†, Michal Rosen-Zvi† , Gustavo Stolovitzky† , Mahtab Mirmomeni† , Stefan Harrer† * These authors contributed equally to this work † Corresponding authors: rkhalaf@us.ibm.com, rosen@il.ibm.com, gustavo@us.ibm.com, mahtabm@au1.ibm.com, sharrer@au.ibm.com ◊ Members of the IBM Epilepsy Consortium are listed in the Acknowledgements section J. Picone and I. Obeid are with Temple University, USA. T. Schaffter is with Sage Bionetworks, USA. E. Mehmet is with the University of Illinois at Urbana-Champaign, USA. All other authors are with IBM Research in USA, Israel and Australia. Introduction This decade has seen an ever-growing number of scientific fields benefitting from the advances in machine learning technology and tooling. More recently, this trend reached the medical domain, with applications reaching from cancer diagnosis [1] to the development of brain-machine-interfaces [2]. While Kaggle has pioneered the crowd-sourcing of machine learning challenges to incentivise data scientists from around the world to advance algorithm and model design, the increasing complexity of problem statements demands of participants to be expert data scientists, deeply knowledgeable in at least one other scientific domain, and competent software engineers with access to large compute resources. People who match this description are few and far between, unfortunately leading to a shrinking pool of possible participants and a loss of experts dedicating their time to solving important problems. Participation is even further restricted in the context of any challenge run on confidential use cases or with sensitive data. Recently, we designed and ran a deep learning challenge to crowd-source the development of an automated labelling system for brain recordings, aiming to advance epilepsy research. A focus of this challenge, run internally in IBM, was the development of a platform that lowers the barrier of entry and therefore mitigates the risk of excluding interested parties from participating. The challenge: enabling wide participation With the goal to run a challenge that mobilises the largest possible pool of participants from IBM (global), we designed a use case around previous work in epileptic seizure prediction [3]. In this “Deep Learning Epilepsy Detection Challenge”, participants were asked to develop an automatic labelling system to reduce the time a clinician would need to diagnose patients with epilepsy. Labelled training and blind validation data for the challenge were generously provided by Temple University Hospital (TUH) [4]. TUH also devised a novel scoring metric for the detection of seizures that was used as basis for algorithm evaluation [5]. In order to provide an experience with a low barrier of entry, we designed a generalisable challenge platform under the following principles: 1. No participant should need to have in-depth knowledge of the specific domain. (i.e. no participant should need to be a neuroscientist or epileptologist.) 2. No participant should need to be an expert data scientist. 3. No participant should need more than basic programming knowledge. (i.e. no participant should need to learn how to process fringe data formats and stream data efficiently.) 4. No participant should need to provide their own computing resources. In addition to the above, our platform should further • guide participants through the entire process from sign-up to model submission, • facilitate collaboration, and • provide instant feedback to the participants through data visualisation and intermediate online leaderboards. The platform The architecture of the platform that was designed and developed is shown in Figure 1. The entire system consists of a number of interacting components. (1) A web portal serves as the entry point to challenge participation, providing challenge information, such as timelines and challenge rules, and scientific background. The portal also facilitated the formation of teams and provided participants with an intermediate leaderboard of submitted results and a final leaderboard at the end of the challenge. (2) IBM Watson Studio [6] is the umbrella term for a number of services offered by IBM. Upon creation of a user account through the web portal, an IBM Watson Studio account was automatically created for each participant that allowed users access to IBM's Data Science Experience (DSX), the analytics engine Watson Machine Learning (WML), and IBM's Cloud Object Storage (COS) [7], all of which will be described in more detail in further sections. (3) The user interface and starter kit were hosted on IBM's Data Science Experience platform (DSX) and formed the main component for designing and testing models during the challenge. DSX allows for real-time collaboration on shared notebooks between team members. A starter kit in the form of a Python notebook, supporting the popular deep learning libraries TensorFLow [8] and PyTorch [9], was provided to all teams to guide them through the challenge process. Upon instantiation, the starter kit loaded necessary python libraries and custom functions for the invisible integration with COS and WML. In dedicated spots in the notebook, participants could write custom pre-processing code, machine learning models, and post-processing algorithms. The starter kit provided instant feedback about participants' custom routines through data visualisations. Using the notebook only, teams were able to run the code on WML, making use of a compute cluster of IBM's resources. The starter kit also enabled submission of the final code to a data storage to which only the challenge team had access. (4) Watson Machine Learning provided access to shared compute resources (GPUs). Code was bundled up automatically in the starter kit and deployed to and run on WML. WML in turn had access to shared storage from which it requested recorded data and to which it stored the participant's code and trained models. (5) IBM's Cloud Object Storage held the data for this challenge. Using the starter kit, participants could investigate their results as well as data samples in order to better design custom algorithms. (6) Utility Functions were loaded into the starter kit at instantiation. This set of functions included code to pre-process data into a more common format, to optimise streaming through the use of the NutsFlow and NutsML libraries [10], and to provide seamless access to the all IBM services used. Not captured in the diagram is the final code evaluation, which was conducted in an automated way as soon as code was submitted though the starter kit, minimising the burden on the challenge organising team. Figure 1: High-level architecture of the challenge platform Measuring success The competitive phase of the "Deep Learning Epilepsy Detection Challenge" ran for 6 months. Twenty-five teams, with a total number of 87 scientists and software engineers from 14 global locations participated. All participants made use of the starter kit we provided and ran algorithms on IBM's infrastructure WML. Seven teams persisted until the end of the challenge and submitted final solutions. The best performing solutions reached seizure detection performances which allow to reduce hundred-fold the time eliptologists need to annotate continuous EEG recordings. Thus, we expect the developed algorithms to aid in the diagnosis of epilepsy by significantly shortening manual labelling time. Detailed results are currently in preparation for publication. Equally important to solving the scientific challenge, however, was to understand whether we managed to encourage participation from non-expert data scientists. Figure 2: Primary occupation as reported by challenge participants Out of the 40 participants for whom we have occupational information, 23 reported Data Science or AI as their main job description, 11 reported being a Software Engineer, and 2 people had expertise in Neuroscience. Figure 2 shows that participants had a variety of specialisations, including some that are in no way related to data science, software engineering, or neuroscience. No participant had deep knowledge and experience in data science, software engineering and neuroscience. Conclusion Given the growing complexity of data science problems and increasing dataset sizes, in order to solve these problems, it is imperative to enable collaboration between people with differences in expertise with a focus on inclusiveness and having a low barrier of entry. We designed, implemented, and tested a challenge platform to address exactly this. Using our platform, we ran a deep-learning challenge for epileptic seizure detection. 87 IBM employees from several business units including but not limited to IBM Research with a variety of skills, including sales and design, participated in this highly technical challenge.more » « less
