Russian Federation

Many benchmarks and challenges for AI and AGI exist, which help to reveal both short- and long-term topics and directions of research. We analyze elementary school Olympiad math tasks as a possible benchmark for AGI that can occupy a certain free niche capturing some limitations of the existing neural and symbolic systems better than other existing both language understanding and mathematical tests. A detailed comparison and analysis of the implications of AGI are provided.

AGI-20

Analyzing Elementary School Olympiad Math Tasks as a Benchmark for AGI

math tasks

ai evaluation

language understanding

The AGI conference series is the only major conference series devoted wholly and specifically to the creation of AI systems possessing general intelligence at the human level and ultimately beyond. By gathering together active researchers in the field, for presentation of results and discussion of ideas, we accelerate our progress toward our common goal.

Underline hosted the two main days of AGI 2020 on Wednesday June 24th and Thursday June 25th. Both days were **live-streamed** and you can read the full program bellow. All presentations are available for a **asynchronous viewing**. All presenters are grouped into panels. You can also hear guest talks by **Ben Goertzel** (SingularityNET), **Valery Tarasov** (Semenov Institute of Chemical Physics), **Tatiana Shavrina** (Sberbank) and **Joscha Bach** (MIT Media Lab). 

The AGI conference series is the only major conference series devoted wholly and specifically to the creation of AI systems possessing general intelligence at the human level and ultimately beyond. By gathering together active researchers in the field, for presentation of results and discussion of ideas, we accelerate our progress toward our common goal.

technical paper

This article offers comprehensive criticism of the Turing test and develops quality criteria for new artificial general intelligence (AGI) as-sessment tests. It is shown that the prerequisites A. Turing drew upon when reducing personality and human consciousness to “suitable branches of thought” reflected the engineering level of his time. In fact, the Turing “imi-tation game” employed only symbolic communication and ignored the physical world. This paper suggests that by restricting thinking ability to symbolic systems alone Turing unknowingly constructed “the wall” that excludes any possibility of transition from a complex observable phenom-enon to an abstract image or concept. It is, therefore, sensible to factor in new requirements for AI (artificial intelligence) maturity assessment when approaching the Turing test. Such AI must support all forms of communica-tion with a human being, and it should be able to comprehend abstract im-ages and specify concepts as well as participate in social practices.

Post-Turing Methodology: Breaking the Wall on the Way to Artificial General Intelligence

This work is devoted to unresolved problems of Artificial General Intelligence is the inefficiency of transfer learning. One of the mechanisms that are used to solve this problem in the area of reinforcement learning is a model-based approach. In the paper we are expanding the schema networks method which allows to extract the logical relationships between objects and actions from the environment data. We present algorithms for training a Delta Schema Network (DSN), predicting future states of the environment and planning actions that will lead to positive reward. DSN shows strong performance of transfer learning on the classic Atari game environment.

Delta Schema Network in Model-based Reinforcement Learning

A novel approach to automated learning of syntactic rules governing natural languages is proposed, based on using probabilities assigned to sentences (and potentially longer word sequences) by trans-former neural network language models to guide symbolic learning processes like clustering and rule induction. This method exploits the learned linguistic knowledge in transformers, without any reference to their inner representations; hence, the technique is readily adaptable to the continuous appearance of more powerful language models.

Guiding Symbolic Natural Language Grammar Induction via Transformer-Based Sequence Probabilities

"The complex socio-technological debate underlying safety critical and ethically relevant issues pertaining to AI development and deployment extends across heterogeneous research subfields and involves in part conflicting positions. In this context, it seems expedient to generate a minimalistic joint transdisciplinary basis disambiguating the references to specific subtypes of AI properties and risks for an error-correction in the transmission of ideas. In this paper, we introduce a high-level trans-disciplinary system clustering of ethical distinction between antithetical
clusters of Type I and Type II systems which extends a cybersecurity oriented AI safety taxonomy with considerations from psychology. Moreover, we review relevant Type I AI risks, reflect upon possible epistemological origins of hypothetical Type II AI from a cognitive sciences perspective and discuss the related human moral perception. Strikingly, our nuanced transdisciplinary analysis yields the figurative formulation of the so-called AI safety paradox identifying AI control and value alignment as conjugate requirements in AI safety. Against this backdrop, we craft
versatile multidisciplinary recommendations with ethical dimensions tailored to Type II AI safety. Overall, we suggest proactive and importantly corrective instead of prohibitive methods as common basis for both Type I and Type II AI safety."

Error-Correction for AI Safety

The videogame platform Teleport created earlier allows us to study anonymous social interactions among actors of various nature: human and virtual actor, ensuring their indistinguishability, which implies believable behavior of a virtual actor. The present study found a connection between the human player behavior in the Teleport game and her psychological portrait constructed using the sixteen-factor Catell personality test for empathy. Assuming that this connection is extendable to perception of virtual actor behavior, the game sessions data was analyzed to infer behavioral characteristics of virtual actors. Based on this data analysis, we constructed psychological characteristics of models of a virtual actor (a bot). Partner and emotional characteristics of bots were defined, and their psychological portrait was constructed based on the registered bot behavior. Personal characteristics such as courage, sociability, calmness, balance, and loyalty were attributed to bots and compared to analogous characteristics of human players.

Psychological Portrait of a Virtual Agent in the Teleport Game Paradigm

"No fundamental new ideas have appeared in AI for decades because of a deadlocked discussion between the technologists and their philosophical critics. Both sides claim possession of the one (dogmatic) truth: Technologists are committed to writing code, while critics insist that AI bears no resemblance to how humans cope in the world. The book charts a middle course between the critics and practitioners of AI, remaining committed to writing code while maintaining a fixed gaze on the human condition. This is done by reviving a technique long-shunned in cognitive science: Introspection. Introspection was rejected as a scientific method since 1913, but technology is committed to “what works” rather than to science’s “best explanation”. Introspection is shown to be both a legitimate and a promising source of ideas for AI. The book details the development process of AI based on introspection, from the initial introspective descriptions to working code.
This book is unique in that it starts with philosophical (and historical) discussions, and ends with examples of working novel algorithms. The book was originally a PhD thesis. It was edited for book form with two new chapters added."

Report on “AI and Human Thought and Emotion”

Central scholars in AI have argued for extending the search for new AI technology beyond the tried-and-tested biologically and mathematically-inspired algorithms. Following in their footsteps, areas in the humanities are introduced as possible inspirations for novel human-like AI. Topics discussed include play-acting, literature as the field researching both imagination and metaphors, linguistics, music, and hermeneutics. In our ambition to reach general intelligence, we cannot afford to ignore these avenues of research.

AGI Needs the Humaities

Social virtual actors need to interact with users emotionally, convincing them in their ability to understand human minds. For this to happen, an artificial general emotional intelligence is needed, capable of believable behavior in real-life situations. Summarizing recent work of the authors, the present work extends the general state-of-the-art framework of emotional AGI, using the emotional Biologically Inspired Cognitive Architecture (eBICA) as a basis. In addition to appraisals, other kinds of fluents are added to the model, including somatic markers, feelings, emotional biases, moods, etc. Their integration is achieved based on semantic maps and moral schemas. It is anticipated that the proposed new level of artificial general socially emotional intelligence will complement the next-generation AGI, helping it to merge into our society on equal with its human members.

Toward a General Believable Model of Human-Analogous Intelligent Socially Emotional Behavior

The DeepWalk algorithm is used to assign embedding vec- tors to nodes in the Atomspace weighted, labeled hypergraph that is used to represent knowledge in the OpenCog AGI system, in the con- text of an application to probabilistic inference regarding the causes of longevity based on data from biological ontologies and genomic analyses. It is shown that vector difference operations between embedding vectors are, in appropriate conditions, approximately alignable with “intensional difference” operations between the hypergraph nodes corresponding to the embedding vectors. This relationship hints at a broader functorial mapping between uncertain intensional logic and vector arithmetic, and opens the door for using embedding vector algebra to guide intensional inference control.

Embedding Vector Differences Can Be Aligned With Uncertain Intensional Logic Differences

What kind of programming language would be most appropriate to serve the needs of integrative, multi-paradigm, multi-software-system approaches to AGI? This question is broached via exploring the more particular question of how to create a more scalable and usable version of the ”Atomese” programming language that forms a key component of the OpenCog AGI design (an ”Atomese 2.0”) .

Analyzing Elementary School Olympiad Math Tasks as a Benchmark for AGI

Downloads

Next from AGI-20

Post-Turing Methodology: Breaking the Wall on the Way to Artificial General Intelligence

Similar lecture

How do you test the strength of AI?

Stay up to date with the latest Underline news!

PRODUCT

COMPANY

RESOURCES

Analyzing Elementary School Olympiad Math Tasks as a Benchmark for AGI

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AGI-20

Post-Turing Methodology: Breaking the Wall on the Way to Artificial General Intelligence

Similar lecture

How do you test the strength of AI?

Stay up to date with the latest Underline news!

PRODUCT

COMPANY

RESOURCES

Downloads