Kirill Krinkin

Kirill Krinkin, Yulia Shichkina

This paper questions the feasibility of a strong (general) data-centric artificial intelligence (AI). The disadvantages of this type of intelligence are discussed. As an alternative, the concept of co-evolutionary hybrid intelligence is proposed. It is based on the cognitive interoperability of man and machine. An analysis of existing approaches to the construction of cognitive architectures is given. An architecture seamlessly incorporates a human into the loop of intelligent problem solving is considered. The article is organized as follows. The first part contains a critique of data-centric intelligent systems. The reasons why it is impossible to create a strong artificial intelligence based on this type of intelligence are indicated. The second part briefly presents the concept of co-evolutionary hybrid intelligence and shows its advantages. The third part gives an overview and analysis of existing cognitive architectures. It is concluded that many do not consider humans part of the intelligent data processing process. The next part discusses the cognitive architecture for co-evolutionary hybrid intelligence, providing integration with humans. It finishes with general conclusions about the feasibility of developing intelligent systems with humans in the problem-solving loop.

Kirill Krinkin

The observation that optimum circuit size changes by at most $O(n)$ under a one-point truth table perturbation is implicit in prior work on the Minimum Circuit Size Problem. This note states the bound explicitly for arbitrary fixed finite complete bases with unit-cost gates, extends it to general Hamming distance via a telescoping argument, and verifies it exhaustively at $n = 4$ in the AIG basis using SAT-derived exact circuit sizes for 220 of 222 NPN equivalence classes. Among 987 mutation edges, the maximum observed difference is $4 = n$, confirming the bound is tight at $n = 4$ for AIG.

Kirill Krinkin

We study the gap between the minimum size of a Boolean circuit (DAG) and the minimum size of a formula (tree circuit) over the And-Inverter Graph (AIG) basis {AND, NOT} with free inversions. We prove that this gap is always 0 or 1 (Unit Gap Theorem), that sharing requires opt(f) >= n essential variables (Threshold Theorem), and that no sharing is needed when opt(f) <= 3 (Tree Theorem). Gate counts in optimal circuits satisfy an exact decomposition formula with a binary sharing term. When the gap equals 1, it arises from exactly one gate with fan-out 2, employing either dual-polarity or same-polarity reuse; we prove that no other sharing structure can produce a unit gap.

Kirill Krinkin, Yulia Shichkina, Andrey Ignatyev

Artificial intelligence is one of the drivers of modern technological development. The current approach to the development of intelligent systems is data-centric. It has several limitations: it is fundamentally impossible to collect data for modeling complex objects and processes; training neural networks requires huge computational and energy resources; solutions are not explainable. The article discusses an alternative approach to the development of artificial intelligence systems based on human-machine hybridization and their co-evolution.

Kirill Krinkin, Anton Filatov

Modern laser SLAM (simultaneous localization and mapping) and structure from motion algorithms face the problem of processing redundant data. Even if a sensor does not move, it still continues to capture scans that should be processed. This paper presents the novel filter that allows dropping 2D scans that bring no new information to the system. Experiments on MIT and TUM datasets show that it is possible to drop more than half of the scans. Moreover thepaper describes the formulas that enable filter adaptation to a particular robot with known speed and characteristics of lidar. In addition, the indoor corridor detector is introduced that also can be applied to any specific shape of a corridor and sensor.

Kirill Krinkin, Igor Dronnikov

Currently, tens of millions of devices around the world communicate with/ each other via cellular networks. In this paper, we study the stability of network content delivery protocols to the effects of network interference. To conduct the research, a tool was developed that allows testing of protocols, such as TCP, UDP, and QUIC. The analysis and comparison of the obtained test results were carried out. In the conclusion, the best protocols for content delivery were shown

Mikita Sazanovich, Konstantin Chaika, Kirill Krinkin et al.

In this paper, we describe our winning approach to solving the Lane Following Challenge at the AI Driving Olympics Competition through imitation learning on a mixed set of simulation and real-world data. AI Driving Olympics is a two-stage competition: at stage one, algorithms compete in a simulated environment with the best ones advancing to a real-world final. One of the main problems that participants encounter during the competition is that algorithms trained for the best performance in simulated environments do not hold up in a real-world environment and vice versa. Classic control algorithms also do not translate well between tasks since most of them have to be tuned to specific driving conditions such as lighting, road type, camera position, etc. To overcome this problem, we employed the imitation learning algorithm and trained it on a dataset collected from sources both from simulation and real-world, forcing our model to perform equally well in all environments.

Anton Filatov, Artyom Filatov, Kirill Krinkin et al.

SLAM (Simultaneous Localization and mapping) is one of the most challenging problems for mobile platforms and there is a huge amount of modern SLAM algorithms. The choice of the algorithm that might be used in every particular problem requires prior knowledge about advantages and disadvantages of each algorithm. This paper presents the approach for comparison of SLAM algorithms that allows to find the most accurate one. The accent of research is made on 2D SLAM algorithms and the focus of analysis is 2D map that is built after algorithm performance. Three metrics for evaluation of maps are presented in this paper