Large language models (LLMs) are known to exhibit biases in downstream tasks, especially when dealing with sensitive topics such as political discourse, gender identity, ethnic relations, or national stereotypes. Although significant progress has been made in bias detection and mitigation techniques, certain challenges remain underexplored. This study proposes a reusable, granular and topic-agnostic framework to evaluate polarisation-related biases in LLM (both open-source and closed-source). Our approach combines polarisation-sensitive sentiment metrics with a synthetically generated balanced dataset of conflict-related statements, using a predefined set of semantic categories. Using our synthetic dataset that focuses on the Russia-Ukraine war, we evaluated the entity-asymmetry in sentiment scoring in six LLMs: Llama-3.1-8B, Ministral-3-14B, Gemma-3-27B, Qwen3-30B-A3B, Claude Sonnet 4.6, and GPT-5.2. Beyond aggregate asymmetries scores, with a general trend for more positive sentiment toward Ukraine, the framework allowed fine-grained analysis with considerable variation between semantic categories, uncovering divergent behavioural patterns among models. Adaptation to prompt modifications showed further entity-asymmetry in sentiment scoring towards preconceived language and citizenship modification. Overall, the framework supports automated dataset generation and fine-grained entity-asymmetry in sentiment scoring assessment, is applicable to a variety of polarisation-driven scenarios and topics, and is complementary to embedding-based and probability-based bias-evaluation strategies.

Abstract Spiking neural P systems (SN P systems) are a mathematical model of neural networks, abstracting the way biological neurons communicate with spikes, developed within the framework of the membrane computing theory. Recently, driven by the boom of learning neural models, SN P systems have become a rapidly emerging research front. Consequently, many different variants of the learning models of SN P system prevail among the new research results. Although large proprietary deep learning models are still based on the continuous neural network paradigm, spiking neurons are attractive because of their low-energy demands. The purpose of this paper is to provide an up-to-date overview of learning paradigms and techniques for SN P systems. After a brief introduction of the structure and function of SN P systems, we summarise recent approaches to learning and adaptation in SN P systems, including Hebbian learning, Widrow-Hoff algorithm, fuzzy approaches, nonlinear SN P systems, gated and long short-term memory inspired SN P systems, convolutional SN P systems, and more.

Success in synthetic biology depends on the efficient construction of robust genetic circuitry. However, even the direct engineering of the simplest genetic elements (switches, logic gates) is a challenge and involves intense lab work. As the complexity of biological circuits grows, it becomes more complicated and less fruitful to rely on the rational design paradigm, because it demands many time-consuming trial-and-error cycles. One of the reasons is the context-dependent behavior of small assembly parts (like BioBricks), which in a complex environment often interact in an unpredictable way. Therefore, the idea of evolutionary engineering (artificial directed in vivo evolution) based on screening and selection of randomized combinatorial genetic circuit libraries became popular. In this article we build on the so-called dual selection technique. We propose a plasmid-based framework using toxin-antitoxin pairs together with the relaxase conjugative protein, enabling an efficient autonomous in vivo evolutionary selection of simple Boolean circuits in bacteria (E. coli was chosen for demonstration). Unlike previously reported protocols, both on and off selection steps can run simultaneously in various cells in the same environment without human intervention; and good circuits not only survive the selection process but are also horizontally transferred by conjugation to the neighbor cells to accelerate the convergence rate of the selection process. Our directed evolution strategy combines a new dual selection method with fluorescence-based screening to increase the robustness of the technique against mutations. As there are more orthogonal toxin-antitoxin pairs in E. coli, the approach is likely to be scalable to more complex functions. In silico experiments based on empirical data confirm the high search and selection capability of the protocol.

A membrane system (P system) is a distributed computing model inspired by information processes in living cells. P systems previously provided new characterizations of a variety of complexity classes and their borderlines. Specifically, in tissue-like membrane systems, cell separation rules have been considered joint with communication rules of the form symport/antiport. On the one hand, only tractable problems can be efficiently solved by using cell separation and communication rules with length at most 2. On the other hand, an efficient and uniform solution to the SAT problem by using cell separation and communication rules with length at most 8 has been recently given. In this paper we improve the previous result by showing that the SAT problem can be solved by a family of tissue P systems with cell separation in linear time, by using communication rules with length at most 3. Thus, in the framework of tissue P systems with cell separation, we provide an optimal tractability borderline: passing from length 2 to 3 amounts to passing from non–efficiency to efficiency, assuming that P ≠ NP.

We improve and extend a recent result showing that spiking neural P systems with the same rules in all neurons of the system (homogenous) and working in the max sequential manner are universal. The previous work in this area reported by the group led

Self-assembly, the process by which objects autonomously come together to form complex structures, is omnipresent in the physical world. Recent experiments in self-assembly demonstrate its potential for the parallel creation of a large number of nanostructures, including possibly computers. A systematic study of self-assembly as a mathematical process has been initiated by L. Adleman and E. Winfree. The individual components are modeled as square tiles on the infinite two-dimensional plane. Each side of a tile is covered by a specific “glue,” and two adjacent tiles will stick iff they have matching glues on their abutting edges. Tiles that stick to each other may form various two-dimensional “structures” such as squares and rectangles, or may cover the entire plane. In this paper we focus on a special type of structure, called a ribbon: a non-self-crossing rectilinear sequence of tiles on the plane, in which successive tiles are adjacent along an edge and abutting edges of consecutive tiles have matching glues. We prove that it is undecidable whether an arbitrary finite set of tiles with glues (infinite supply of each tile type available) can be used to assemble an infinite ribbon. While the problem can be proved undecidable using existing techniques if the ribbon is required to start with a given “seed” tile, our result settles the “unseeded” case, an open problem formerly known as the “unlimited infinite snake problem.” The proof is based on a construction, due to R. Robinson, of a special set of tiles that allow only aperiodic tilings of the plane. This construction is used to create a special set of directed tiles (tiles with arrows painted on the top) with the “strong plane-filling property”—a variation of the “plane-filling property” previously defined by J. Kari. A construction of “sandwich” tiles is then used in conjunction with this special tile set, to reduce the well-known undecidable tiling problem to the problem of the existence of an infinite directed zipper (a special kind of ribbon). A “motif” construction is then introduced that allows one tile system to simulate another by using geometry to represent glues. Using motifs, the infinite directed zipper problem is reduced to the infinite ribbon problem, proving the latter undecidable. An immediate consequence of our result is the undecidability of the existence of arbitrarily large structures self-assembled using tiles from a given tile set.