Artur Hakobyan

Vardan Adibekyan, Olivier Demangeon, Tiago Campante et al.

Citations are essential for recognizing scientific contributions, yet citation behavior is shaped by more than just relevance or quality. We analyzed approximately 255,000 refereed astronomy articles published between 2000 and 2025 to investigate how journals are cited relative to their publication volume and authorship context. We find that multidisciplinary journals receive disproportionately more citations, up to nine times higher than their share of articles, while field-specific journals are cited less frequently in proportion to their output. Citations to a journal also increase significantly when authors publish within it, a bias particularly pronounced in multidisciplinary journals. Although this effect has declined over the past decade, it remains notable. These patterns likely arise from a combination of topical clustering, institutional/individual publishing habits, and strategic referencing to align with editorial expectations. Our findings reveal persistent structural biases in scientific visibility and suggest that citation-based metrics should be used with greater awareness of the publishing context they reflect. We encourage authors, reviewers, and editors to remain mindful of these dynamics and strive for fairness and inclusivity when selecting references.

Khachik Smbatyan, Tsolak Ghukasyan, Tigran Aghajanyan et al.

The rapid advancement of artificial intelligence, particularly autonomous agentic systems based on Large Language Models (LLMs), presents new opportunities to accelerate drug discovery by improving in-silico modeling and reducing dependence on costly experimental trials. Current AI agent-based systems demonstrate proficiency in solving programming challenges and conducting research, indicating an emerging potential to develop software capable of addressing complex problems such as pharmaceutical design and drug discovery. This paper introduces DO Challenge, a benchmark designed to evaluate the decision-making abilities of AI agents in a single, complex problem resembling virtual screening scenarios. The benchmark challenges systems to independently develop, implement, and execute efficient strategies for identifying promising molecular structures from extensive datasets, while navigating chemical space, selecting models, and managing limited resources in a multi-objective context. We also discuss insights from the DO Challenge 2025, a competition based on the proposed benchmark, which showcased diverse strategies explored by human participants. Furthermore, we present the Deep Thought multi-agent system, which demonstrated strong performance on the benchmark, outperforming most human teams. Among the language models tested, Claude 3.7 Sonnet, Gemini 2.5 Pro and o3 performed best in primary agent roles, and GPT-4o, Gemini 2.0 Flash were effective in auxiliary roles. While promising, the system's performance still fell short of expert-designed solutions and showed high instability, highlighting both the potential and current limitations of AI-driven methodologies in transforming drug discovery and broader scientific research.