Towards a Quantum-Like Cognitive Architecture for Decision-Making Artificial Intelligence

We propose an alternative and unifying framework for decision-making that, by using quantum mechanics, provides more generalised cognitive and decision models with the ability to represent more information than classical models. This framework can accommodate and predict several cognitive biases reported in Lieder & Griffiths without heavy reliance on heuristics nor on assumptions of the computational resources of the mind. Expected utility theory and classical probabilities tell us what people should do if employing traditionally rational thought, but do not tell us what people do in reality (Machina, 2009). Under this principle, L&G propose an architecture for cognition that can serve as an intermediary layer between Neuroscience and Computation. Whilst instances where large expenditures of cognitive resources occur are theoretically alluded to, the model primarily assumes a preference for fast, heuristic-based processing.

Transcriptome and epigenome landscape of human cortical development modeled in organoids


The human cerebral cortex has undergone an extraordinary increase in size and complexity during mammalian evolution. Cortical cell lineages are specified in the embryo, and genetic and epidemiological evidence implicates early cortical development in the etiology of neuropsychiatric disorders such as autism spectrum disorder (ASD), intellectual disabilities, and schizophrenia. Most of the disease-implicated genomic variants are located outside of genes, and the interpretation of noncoding mutations is lagging behind owing to limited annotation of functional elements in the noncoding genome. We set out to discover gene-regulatory elements and chart their dynamic activity during prenatal human cortical development, focusing on enhancers, which carry most of the weight upon regulation of gene expression. We longitudinally modeled human brain development using human induced pluripotent stem cell (hiPSC)–derived cortical organoids and compared organoids to isogenic fetal brain tissue. Fetal fibroblast–derived hiPSC lines were used to generate cortically patterned organoids and to compare oganoids' epigenome and transcriptome to that of isogenic fetal brains and external datasets. Organoids model cortical development between 5 and 16 postconception weeks, thus enabling us to study transitions from cortical stem cells to progenitors to early neurons. The greatest changes occur at the transition from stem cells to progenitors. The regulatory landscape encompasses a total set of 96,375 enhancers linked to target genes, with 49,640 enhancers being active in organoids but not in mid-fetal brain, suggesting major roles in cortical neuron specification. Enhancers that gained activity in the human lineage are active in the earliest stages of organoid development, when they target genes that regulate the growth of radial glial cells. Parallel weighted gene coexpression network analysis (WGCNA) of transcriptome and enhancer activities defined a number of modules of coexpressed genes and coactive enhancers, following just six and four global temporal patterns that we refer to as supermodules, likely reflecting fundamental programs in embryonic and fetal brain. Correlations between gene expression and enhancer activity allowed stratifying enhancers into two categories: activating regulators (A-regs) and repressive regulators (R-regs).

Companion-Based Ambient Robust Intelligence (CARING)

AAAI Conferences

We present a Companion-based Ambient Robust INtelliGence (CARING) system, for communication with, and support of, clients with Traumatic brain injury (TBI) or Amyotrophic Lateral Sclerosis (ALS). A central component of this system is an artificial companion, combined with a range of elements for ambient intelligence. The companion acts as a personalized intermediary for multi-party communication between the client, the environment (e.g. a Smart Home), caregivers and health professionals. CARING is based on tightly coupled systems drawing from natural language processing, speech recognition and adaptation, deep language understanding and constraint-based knowledge representation and reasoning. A major innovation of the system is its ability to adapt and accommodate different interfaces associated with different client capabilities and needs. The system will use, as a proxy, different interaction requirements of clients (e.g., Brain-Computer Interfaces) at different stages of ALS progression and with different types of TBI impairments. Ultimately, this technology is expected to improve the quality of life for clients through conversation with a computer.

Knowledge Discovery using Genetic Programming with Rough Set Evaluation

AAAI Conferences

An important area of KDD research involves development of techniques which transform raw data into forms more useful for prediction or explanation. We present an approach to automating the search for "indicator functions" which mediate such transformations. The fitness of a function is measured as its contribution to discerning different classes of data. Genetic programming techniques are applied to the search for and improvement of the programs which make up these functions. Rough set theory is used to evaluate the fitness of functions. Rough set theory provides a unique evaluator in that it allows the fitness of each function to depend on the combined performance of a population of functions. This is desirable in applications which need a population of programs that perform well in concert and contrasts with traditional genetic programming applications which have as there goal to find a single program which performs well. This approach has been applied to a small database of iris flowers with the goal of learning to predict the species of the flower given the values of four iris attributes and to a larger breast cancer database with the goal of predicting whether remission will occur within a five year period.


AAAI Conferences

While they are very useful to diagnose typical cases, it is difficult for them to diagnose complicated cases. Therefore various approaches, such as deeper knowledge representation, case-based reasoning, are proposed in order to overcome this problem. However, they axe not sufficient to solve this problem completely. One reason that they are not so sutticient is that they are lacking one important track of diagnosis that medical experts do when they meet complicated cases. In this paper, we introduce combination of reasoning, planning and learning methods in order to solve this difficulty.