NeuroLISP: High-level symbolic programming with attractor neural networks.

Despite significant improvements in contemporary machine learning, symbolic methods currently outperform artificial neural networks on tasks that involve compositional reasoning, such as goal-directed planning and logical inference. This illustrates a computational explanatory gap between cognitive and neurocomputational algorithms that obscures the neurobiological mechanisms underlying cognition and impedes progress toward human-level artificial intelligence. Because of the strong relationship between cognition and working memory control, we suggest that the cognitive abilities of contemporary neural networks are limited by biologically-implausible working memory systems that rely on persistent activity maintenance and/or temporal nonlocality. Here we present NeuroLISP, an attractor neural network that can represent and execute programs written in the LISP programming language. Unlike previous approaches to high-level programming with neural networks, NeuroLISP features a temporally-local working memory based on itinerant attractor dynamics, top-down gating, and fast associative learning, and implements several high-level programming constructs such as compositional data structures, scoped variable binding, and the ability to manipulate and execute programmatic expressions in working memory (i.e., programs can be treated as data). Our computational experiments demonstrate the correctness of the NeuroLISP interpreter, and show that it can learn non-trivial programs that manipulate complex derived data structures (multiway trees), perform compositional string manipulation operations (PCFG SET task), and implement high-level symbolic AI algorithms (first-order unification). We conclude that NeuroLISP is an effective neurocognitive controller that can replace the symbolic components of hybrid models, and serves as a proof of concept for further development of high-level symbolic programming in neural networks.

Compositional memory in attractor neural networks with one-step learning.

Compositionality refers to the ability of an intelligent system to construct models out of reusable parts. This is critical for the productivity and generalization of human reasoning, and is considered a necessary ingredient for human-level artificial intelligence. While traditional symbolic methods have proven effective for modeling compositionality, artificial neural networks struggle to learn systematic rules for encoding generalizable structured models. We suggest that this is due in part to short-term memory that is based on persistent maintenance of activity patterns without fast weight changes. We present a recurrent neural network that encodes structured representations as systems of contextually-gated dynamical attractors called attractor graphs. This network implements a functionally compositional working memory that is manipulated using top-down gating and fast local learning. We evaluate this approach with empirical experiments on storage and retrieval of graph-based data structures, as well as an automated hierarchical planning task. Our results demonstrate that compositional structures can be stored in and retrieved from neural working memory without persistent maintenance of multiple activity patterns. Further, memory capacity is improved by the use of a fast store-erase learning rule that permits controlled erasure and mutation of previously learned associations. We conclude that the combination of top-down gating and fast associative learning provides recurrent neural networks with a robust functional mechanism for compositional working memory.

Tunable Neural Encoding of a Symbolic Robotic Manipulation Algorithm.

We present a neurocomputational controller for robotic manipulation based on the recently developed "neural virtual machine" (NVM). The NVM is a purely neural recurrent architecture that emulates a Turing-complete, purely symbolic virtual machine. We program the NVM with a symbolic algorithm that solves blocks-world restacking problems, and execute it in a robotic simulation environment. Our results show that the NVM-based controller can faithfully replicate the execution traces and performance levels of a traditional non-neural program executing the same restacking procedure. Moreover, after programming the NVM, the neurocomputational encodings of symbolic block stacking knowledge can be fine-tuned to further improve performance, by applying reinforcement learning to the underlying neural architecture.

A programmable neural virtual machine based on a fast store-erase learning rule.

We present a neural architecture that uses a novel local learning rule to represent and execute arbitrary, symbolic programs written in a conventional assembly-like language. This Neural Virtual Machine (NVM) is purely neurocomputational but supports all of the key functionality of a traditional computer architecture. Unlike other programmable neural networks, the NVM uses principles such as fast non-iterative local learning, distributed representation of information, program-independent circuitry, itinerant attractor dynamics, and multiplicative gating for both activity and plasticity. We present the NVM in detail, theoretically analyze its properties, and conduct empirical computer experiments that quantify its performance and demonstrate that it works effectively.

Using Directional Fibers to Locate Fixed Points of Recurrent Neural Networks.

We introduce mathematical objects that we call "directional fibers," and show how they enable a new strategy for systematically locating fixed points in recurrent neural networks. We analyze this approach mathematically and use computer experiments to show that it consistently locates many fixed points in many networks with arbitrary sizes and unconstrained connection weights. Comparison with a traditional method shows that our strategy is competitive and complementary, often finding larger and distinct sets of fixed points. We provide theoretical groundwork for further analysis and suggest next steps for developing the method into a more powerful solver.

Humanoid Cognitive Robots That Learn by Imitating: Implications for Consciousness Studies.

While the concept of a conscious machine is intriguing, producing such a machine remains controversial and challenging. Here, we describe how our work on creating a humanoid cognitive robot that learns to perform tasks via imitation learning relates to this issue. Our discussion is divided into three parts. First, we summarize our previous framework for advancing the understanding of the nature of phenomenal consciousness. This framework is based on identifying computational correlates of consciousness. Second, we describe a cognitive robotic system that we recently developed that learns to perform tasks by imitating human-provided demonstrations. This humanoid robot uses cause-effect reasoning to infer a demonstrator's intentions in performing a task, rather than just imitating the observed actions verbatim. In particular, its cognitive components center on top-down control of a working memory that retains the explanatory interpretations that the robot constructs during learning. Finally, we describe our ongoing work that is focused on converting our robot's imitation learning cognitive system into purely neurocomputational form, including both its low-level cognitive neuromotor components, its use of working memory, and its causal reasoning mechanisms. Based on our initial results, we argue that the top-down cognitive control of working memory, and in particular its gating mechanisms, is an important potential computational correlate of consciousness in humanoid robots. We conclude that developing high-level neurocognitive control systems for cognitive robots and using them to search for computational correlates of consciousness provides an important approach to advancing our understanding of consciousness, and that it provides a credible and achievable route to ultimately developing a phenomenally conscious machine.

A limit-cycle self-organizing map architecture for stable arm control.

Inspired by the oscillatory nature of cerebral cortex activity, we recently proposed and studied self-organizing maps (SOMs) based on limit cycle neural activity in an attempt to improve the information efficiency and robustness of conventional single-node, single-pattern representations. Here we explore for the first time the use of limit cycle SOMs to build a neural architecture that controls a robotic arm by solving inverse kinematics in reach-and-hold tasks. This multi-map architecture integrates open-loop and closed-loop controls that learn to self-organize oscillatory neural representations and to harness non-fixed-point neural activity even for fixed-point arm reaching tasks. We show through computer simulations that our architecture generalizes well, achieves accurate, fast, and smooth arm movements, and is robust in the face of arm perturbations, map damage, and variations of internal timing parameters controlling the flow of activity. A robotic implementation is evaluated successfully without further training, demonstrating for the first time that limit cycle maps can control a physical robot arm. We conclude that architectures based on limit cycle maps can be organized to function effectively as neural controllers.

Towards a multi-level neural architecture that unifies self-intended and imitated arm reaching performance.

Dexterous arm reaching movements are a critical feature that allow human interactions with tools, the environment, and socially with others. Thus the development of a neural architecture providing unified mechanisms for actual, mental, observed and imitated actions could enhance robot performance, enhance human-robot social interactions, and inform specific human brain processes. Here we present a model, including a fronto-parietal network that implements sensorimotor transformations (inverse kinematics, workspace visuo-spatial rotations), for self-intended and imitation performance. Our findings revealed that this neural model can perform accurate and robust 3D actual/mental arm reaching while reproducing human-like kinematics. Also, using visuo-spatial remapping, the neural model can imitate arm reaching independently of a demonstrator-imitator viewpoint. This work is a first step towards providing the basis of a future neural architecture for combining cognitive and sensorimotor processing levels that will allow for multi-level mental simulation when executing actual, mental, observed, and imitated actions for dexterous arm movements.

Morphology of influenza B/Lee/40 determined by cryo-electron microscopy.

Cryo-electron microscopy projection image analysis and tomography is used to describe the overall architecture of influenza B/Lee/40. Algebraic reconstruction techniques with utilization of volume elements (blobs) are employed to reconstruct tomograms of this pleomorphic virus and distinguish viral surface spikes. The purpose of this research is to examine the architecture of influenza type B virions by cryo-electron tomography and projection image analysis. The aims are to explore the degree of ribonucleoprotein disorder in irregular shaped virions; and to quantify the number and distribution of glycoprotein surface spikes (hemagglutinin and neuraminidase) on influenza B. Projection image analysis of virion morphology shows that the majority (∼83%) of virions are spherical with an average diameter of 134±19 nm. The aspherical virions are larger (average diameter = 155±47 nm), exhibit disruption of the ribonucleoproteins, and show a partial loss of surface protein spikes. A count of glycoprotein spikes indicates that a typical 130 nm diameter type B virion contains ∼460 surface spikes. Configuration of the ribonucleoproteins and surface glycoprotein spikes are visualized in tomogram reconstructions and EM densities visualize extensions of the spikes into the matrix. The importance of the viral matrix in organization of virus structure through interaction with the ribonucleoproteins and the anchoring of the glycoprotein spikes to the matrix is demonstrated.

Bacteriophage φ6--structure investigated by fluorescence Stokes shift spectroscopy.

The Stokes shift of tryptophan (Trp) fluorescence from layers of the lipid-containing bacteriophage φ6 is compared to determine the relative effect of the layers on virus hydrophobicity. In the inner most layer, the empty procapsid (PC) which contains 80-90% of the virion Trp residues, λ(max) = 339.8 nm. The PC emission is substantially more redshifted than the other φ6 layers and nearer to that of the Pseudomonad host cell than the other φ6 layers. The Trp emission from the nucleocapsid (NC) with λ(max) = 337.4 nm, is blueshifted by 2.4 nm relative to the PC although the number of Trp in the NC is identical to the PC. This shift represents an increase in Trp hydrophobicity, likely a requirement for the maintenance of A-form doubled-stranded RNA. Fluorescence from the completely assembled virion indicates it is in a considerably more hydrophobic environment with λ(max) = 330.9 nm. Density measurements show that the water content in the NC does not change during envelope assembly, therefore the blueshifted φ6 emission suggests that the envelope changes the PC environment, probably via the P8 layer. This change in hydrophobicity likely arises from charge redistribution or envelope-induced structural changes in the PC proteins.

Protein P7 of the cystovirus φ6 is located at the three-fold axis of the unexpanded procapsid.

The objective of this study was to determine the location of protein P7, the RNA packaging factor, in the procapsid of the φ6 cystovirus. A comparison of cryo-electron microscopy high-resolution single particle reconstructions of the φ6 complete unexpanded procapsid, the protein P2-minus procapsid (P2 is the RNA directed RNA-polymerase), and the P7-minus procapsid, show that prior to RNA packaging the P7 protein is located near the three-fold axis of symmetry. Difference maps highlight the precise position of P7 and demonstrate that in P7-minus particles the P2 proteins are less localized with reduced densities at the three-fold axes. We propose that P7 performs the mechanical function of stabilizing P2 on the inner protein P1 shell which ensures that entering viral single-stranded RNA is replicated.

Toroidal surface complexes of bacteriophage ϕ12 are responsible for host-cell attachment.

Cryo-electron tomography and subtomogram averaging are utilized to determine that the bacteriophage ϕ12, a member of the Cystoviridae family, contains surface complexes that are toroidal in shape, are composed of six globular domains with six-fold symmetry, and have a discrete density connecting them to the virus membrane-envelope surface. The lack of this kind of spike in a reassortant of ϕ12 demonstrates that the gene for the hexameric spike is located in ϕ12's medium length genome segment, likely to the P3 open reading frames which are the proteins involved in viral-host cell attachment. Based on this and on protein mass estimates derived from the obtained averaged structure, it is suggested that each of the globular domains is most likely composed of a total of four copies of P3a and/or P3c proteins. Our findings may have implications in the study of the evolution of the cystovirus species in regard to their host specificity.