In scientific research, collaboration is one of the most effective ways to take advantage of new ideas, skills, resources, and for performing interdisciplinary research. Although collaboration networks have been intensively studied, the question of how individual scientists choose collaborators to study a new research topic remains almost unexplored. Here, we investigate the statistics and mechanisms of collaborations of individual scientists along their careers, revealing that, in general, collaborators are involved in significantly fewer topics than expected from controlled surrogate. In particular, we find that highly productive scientists tend to have higher fraction of single-topic collaborators, while highly cited, i.e., impactful, scientists have higher fraction of multi-topic collaborators. We also suggest a plausible mechanism for this distinction. Moreover, we investigate the cases where scientists involve existing collaborators into a new topic. We find that compared to productive scientists, impactful scientists show strong preference of collaboration with high impact scientists on a new topic. Finally, we validate our findings by investigating active scientists in diff
Extensive research has documented the immediate impacts of the COVID-19 pandemic on scientists, yet it remains unclear if and how such impacts have shifted over time. Here we compare results from two surveys of principal investigators, conducted between April 2020 and January 2021, along with analyses of large-scale publication data. We find that there has been a clear sign of recovery in some regards, as scientists' time spent on their work has almost returned to pre-pandemic levels. However, the latest data also reveals a new dimension in which the pandemic is affecting the scientific workforce: the rate of initiating new research projects. Except for the small fraction of scientists who directly engaged in COVID-related research, most scientists started significantly fewer new research projects in 2020. This decline is most pronounced amongst the same demographic groups of scientists who reported the largest initial disruptions: female scientists and those with young children. Yet in sharp contrast to the earlier phase of the pandemic, when there were large disparities across scientific fields, this loss of new projects appears remarkably homogeneous across fields. Analyses of l
The emergence of Artificial Intelligence (AI) Scientist represents a paradigm shift in scientific discovery, with large language models (LLMs) taking the lead as the primary executor in the entire scientific workflow from idea generation to experiment implementation. Recent AI Scientist studies demonstrate sufficient capabilities for independent scientific discovery, with the generated research reports gaining acceptance at the ICLR 2025 workshop and ACL 2025, arguing that a human-level AI Scientist, capable of uncovering phenomena previously unknown to humans, may be imminent. Despite this substantial progress, AI Scientist has yet to produce a groundbreaking achievement in the domain of computer science on par with automated scientific tools. Based on extensive quantitative evidence from existing benchmarks in complex engineering tasks and a systematic evaluation assess 28 research papers generated by five advanced AI Scientist systems, we argue that \textbf{the fundamental bottleneck for AI Scientists lies in their capability to execute the requisite verification procedures.} Current AI Scientist systems lack the execution capabilities needed to execute rigorous experiments and
The emergence of large language models (LLMs) is propelling automated scientific discovery to the next level, with LLM-based Artificial Intelligence (AI) Scientist systems now taking the lead in scientific research. Several influential works have already appeared in the field of AI Scientist systems, with AI-generated research papers having been accepted at the ICLR 2025 workshop, suggesting that a human-level AI Scientist capable of uncovering phenomena previously unknown to humans, may soon become a reality. In this survey, we focus on the central question: How far are AI scientists from changing the world and reshaping the scientific research paradigm? To answer this question, we provide a prospect-driven review that comprehensively analyzes the current achievements of AI Scientist systems, identifying key bottlenecks and the critical components required for the emergence of a scientific agent capable of producing ground-breaking discoveries that solve grand challenges. We hope this survey will contribute to a clearer understanding of limitations of current AI Scientist systems, showing where we are, what is missing, and what the ultimate goals for scientific AI should be.
Understanding the current capabilities and risks of AI Scientist systems (autoresearch) is essential for ensuring trustworthy and sustainable AI-driven scientific progress while preserving the integrity of the academic ecosystem. To this end, we develop Jr. AI Scientist, a state-of-the-art autonomous AI scientist system that mimics the core research workflow of a novice student researcher: Given the baseline paper from the human mentor, it analyzes its limitations, formulates novel hypotheses for improvement, iteratively experiments until improvements are achieved, and writes a paper with the results. Unlike previous approaches that assume full automation or operate on small-scale code, Jr. AI Scientist follows a well-defined research workflow and leverages modern coding agents to handle complex, multi-file implementations, leading to scientifically valuable contributions. Through our experiments, the Jr. AI Scientist successfully generated new research papers that build upon real NeurIPS, IJCV, and ICLR works by proposing and implementing novel methods. For evaluation, we conducted automated assessments using AI Reviewers, author-led evaluations, and submissions to Agents4Science,
Autonomous systems that generate scientific hypotheses, conduct experiments, and draft manuscripts have recently emerged as a promising paradigm for accelerating discovery. However, existing AI Scientists remain largely domain-agnostic, limiting their applicability to clinical medicine, where research is required to be grounded in medical evidence with specialized data modalities. In this work, we introduce Medical AI Scientist, the first autonomous research framework tailored to clinical autonomous research. It enables clinically grounded ideation by transforming extensively surveyed literature into actionable evidence through clinician-engineer co-reasoning mechanism, which improves the traceability of generated research ideas. It further facilitates evidence-grounded manuscript drafting guided by structured medical compositional conventions and ethical policies. The framework operates under 3 research modes, namely paper-based reproduction, literature-inspired innovation, and task-driven exploration, each corresponding to a distinct level of automated scientific inquiry with progressively increasing autonomy. Comprehensive evaluations by both large language models and human expe
Cyber offense is moving to machine speed; cyber research itself is not. Existing AI scientist systems make end-to-end research automation increasingly plausible, but they target relatively stable scientific domains. We argue that AI-native cybersecurity is a different kind of scientific object. Its recurring units of study are security events and interaction traces, not static assets; its model and tool substrate is non-stationary, not steady-state; and credible evaluation depends on digital twins, cyber ranges, and auditable evidence rather than on a single benchmark score. We call this object the Cybersecurity AI Scientist. A practical realization is a modular, role-specialized multi-agent research system that coordinates problem framing, threat modeling, tool generation, controlled experimentation, evaluation, governance, and scientific reporting, and that anchors its concrete objectives in a four-zeros frame spanning risk, trust, incident, and energy dimensions. As a representative agenda we focus on AI-native defense, where steady-state perimeters give way to resilient agent legions and the classical category of terminal security is itself being deconstructed into agent securi
Product data scientists often ask LLM-based agents to help with recurring execution tasks such as cleaning data, writing SQL, choosing statistical tests, and formatting results. Reusable skill files are meant to avoid prompting from scratch by packaging guidance for a task family. Expert-written skills can encode high-quality guidance, but writing and maintaining them across many data-science task families creates a manual bottleneck. We ask whether LLM-generated skills offer a useful low-curation alternative: do they improve performance over the task prompt alone? We test this question across four lifecycle stages: data preparation, data extraction, statistical analysis, and reporting, using one generated skill per stage. We find no reliable improvement from full generated skills over No-Skill prompting. We then ask whether any part of the skill is useful by ablating different skill components. The main ablation covers 56 tasks, nine model configurations, and three providers, yielding 7,560 runs. Compared with prompting using the task alone, neither the full generated skill nor any ablated skill variant significantly improves performance; all p-values are at least 0.396, and the t
Imagine decision-makers uploading data and, within minutes, receiving clear, actionable insights delivered straight to their fingertips. That is the promise of the AI Data Scientist, an autonomous Agent powered by large language models (LLMs) that closes the gap between evidence and action. Rather than simply writing code or responding to prompts, it reasons through questions, tests ideas, and delivers end-to-end insights at a pace far beyond traditional workflows. Guided by the scientific tenet of the hypothesis, this Agent uncovers explanatory patterns in data, evaluates their statistical significance, and uses them to inform predictive modeling. It then translates these results into recommendations that are both rigorous and accessible. At the core of the AI Data Scientist is a team of specialized LLM Subagents, each responsible for a distinct task such as data cleaning, statistical testing, validation, and plain-language communication. These Subagents write their own code, reason about causality, and identify when additional data is needed to support sound conclusions. Together, they achieve in minutes what might otherwise take days or weeks, enabling a new kind of interaction
Scholars are often categorized into two types: hedgehogs (specialists), who focus on working within a specific research field, and foxes (generalists), who actively contribute to a variety of fields. Despite the familiar anecdotes and popularity of this distinction, its empirical foundation has remained largely unexamined. We examine whether the research style of being a fox or a hedgehog is a stable personal trait or an evolving strategy over a scientist's career. Analyzing 2.3 million scholars' publication records over a century, we find that research styles exhibit remarkable stability. Notably, the proportion of fox-like scientists has dramatically declined in the past century, a phenomenon we term "the death of Renaissance scientists." This decline is particularly significant as science shifts toward team collaboration. Teams of foxes consistently outperform teams of hedgehogs in generating new ideas and directions, as confirmed by two emerging innovation metrics for papers: atypicality and disruption. Our research is the first to quantify the process and consequences of the decline of Renaissance scientists. By doing so, we establish a universal link between research styles,
We present flux measurements of Uranus observed at phase angles of 43.9°, 44.0°, and 52.4° by the Multispectral Visible Imaging Camera (MVIC) on the New Horizons spacecraft during 2023, 2010, and 2019, respectively. New Horizons imaged Uranus at a distance of about 24-70 AU (2023) in four color filters, with bandpasses of 400-550 nm, 540-700 nm, 780-975 nm, and 860-910 nm. High-phase-angle observations are of interest for studying the energy balance of Uranus, constraining the atmospheric scattering behavior, and understanding the planet as an analog for ice giant exoplanets. The new observations from New Horizons provide access to a wider wavelength range and different season compared to previous observations from both Voyager spacecraft. We performed aperture photometry on the New Horizons observations of Uranus to obtain its brightness in each photometric band. The photometry suggests that Uranus may be darker than predicted by a Lambertian phase curve in the Blue and Red filters. Comparison to simultaneous low-phase Hubble WFC3 and ground-based community-led observations indicates a lack of large-scale features at full-phase that would introduce variation in the rotational ligh
Data scientists are not mathematicians, but they make heavy use of mathematics in their daily work. While mathematicians can study a mathematical object which is inaccessible to our five senses, data scientists must deal with real-world data which are observable to us. This fine line suggests that a data scientist's philosophical position on mathematics might have a nontrivial impact on their work. By examining how different philosophical views of mathematics affect the interpretation of the basic model assumption in data science, we arrive at the conclusion that a data scientist, who uses modern probabilistic and statistical tools, must be a Platonist.
We propose a new method to conserve the total energy to round-off error in grid-based codes for hydrodynamic simulations with self-gravity. A formula for the energy flux due to the work done by the the self-gravitational force is given, so the change in total energy can be written in conservative form. Numerical experiments with the code Athena show that the total energy is indeed conserved with our new algorithm and the new algorithm is second order accurate. We have performed a set of tests that show the numerical errors in the traditional, non-conservative algorithm can affect the dynamics of the system. The new algorithm only requires one extra solution of the Poisson equation, as compared to the traditional algorithm which includes self-gravity as a source term. If the Poisson solver takes a negligible fraction of the total simulation time, such as when FFTs are used, the new algorithm is almost as efficient as the original method. This new algorithm is useful in Eulerian hydrodynamic simulations with self-gravity, especially when results are sensitive to small energy errors, as for radiation pressure dominated flow.
In the present article, we present a new gravitational galactic model, describing motion in elliptical as well as in disk galaxies, by suitably choosing the dynamical parameters. Moreover, a new dynamical parameter, the S(g) spectrum, is introduced and used, in order to detect islandic motion of resonant orbits and the evolution of the sticky regions. We investigate the regular or chaotic character of motion, with emphasis in the different dynamical models and make an extensive study of the sticky regions of the system. We use the classical method of the Poincare (r-pr) phase plane and the new dynamical parameter of the S(g) spectrum. The LCE is used, in order to make an estimation of the degree of chaos in our galactic model. In both cases, the numerical calculations, suggest that our new model, displays a wide variety of families of regular orbits, compared to other galactic models. In addition to the regular motion, this new model displays also chaotic regions. Furthermore, the extent of the chaotic regions increases, as the value of the flatness parameter b of the model increases. Moreover, our simulations indicate, that the degree of chaos in elliptical galaxies, is much small
Using the New Horizons LORRI camera, we searched for satellites near five Kuiper belt objects (KBOs): four cold classicals (CCs: 2011 JY31, 2014 OS393, 2014 PN70, 2011 HZ102) and one scattered disk object (SD: 2011 HK103). These objects were observed at distances of 0.092-0.290 au from the New Horizons spacecraft, achieving spatial resolutions of 136-430 km (resolution is ~2 camera pixels), much higher than possible from any other facilities. Here we report that CC 2011 JY31 is a binary system with roughly equal brightness components, CC 2014 OS393 is likely an equal brightness binary system, while the three other KBOs did not show any evidence of binarity. The 2011 JY31 binary has a semi-major axis of 198.6 +/- 2.9 km, an orbital inclination of 61.34 +/- 1.34 deg, and an orbital period of 1.940 +/- 0.002 d. The 2014 OS393 binary objects have an apparent separation of ~150 km, making 2011 JY31 and 2014 OS393 the tightest KBO binary systems ever resolved. Both 2011 HK103 and 2011 HZ102 were detected with SNR~10, and our observations rule out equal brightness binaries with separations larger than ~430 km and ~260 km, respectively. The spatial resolution for 2014 PN70 was ~200 km, but
This paper is a faithful description of the author's career as a scientist, which often intersected that of Yves Couder. The emphasis of this paper is a true description of how the science that the author has been associated with really came about. Included are brief descriptions of the science associated with the research paths described. It is hoped that this rather accurate account may be amusing for the senior scientists among us and educational (and possibly useful) for younger scientists.
In this paper we investigate the opportunities offered by the new Earth gravity models from the dedicated CHAMP and, especially, GRACE missions to the project of measuring the general relativistic Lense-Thirring effect with a new Earth's artificial satellite. It turns out that it would be possible to abandon the stringent, and expensive, requirements on the orbital geometry of the originally prosed LARES mission (same semimajor axis a=12270 km of the existing LAGEOS and inclination i=70 deg) by inserting the new spacecraft in a relatively low, and cheaper, orbit (a=7500-8000 km, i\sim 70 deg) and suitably combining its node Omega with those of LAGEOS and LAGEOS II in order to cancel out the first even zonal harmonic coefficients of the multipolar expansion of the terrestrial gravitational potential J_2, J_4 along with their temporal variations. The total systematic error due to the mismodelling in the remaining even zonal harmonics would amount to \sim 1% and would be insensitive to departures of the inclination from the originally proposed value of many degrees. No semisecular long-period perturbations would be introduced because the period of the node, which is also the period of
One of the most striking and curious features of the small Kuiper Belt Object (KB), Arrokoth, explored by New Horizons, is the bright, annular neck it exhibits at the junction between its two lobes. Here we summarize past reported findings regarding the properties of this feature and report new findings regarding its dimensions, reflectivity and color, shape profile, and its lack of identifiable craters. We conclude by enumerating possible origin scenarios for this unusual feature. New results include a new measurement of the observed neck area of 8+/-1.5 km2, a total neck surface area of 32 km2, a 12.5:1 ratio of neck circumference to height, a normal reflectance histogram of the observed neck, and the fact that no significant (i.e., >2 sigma) color units were identified, meaning the neck's color is generally spatially uniform at the 1.5 km/pixel scale of the best color images. Although several origin hypotheses for the bright material in the neck are briefly discussed, none can be conclusively demonstrated to be the actual origin mechanism at this time; some future tests are identified.
Scientists have uncovered new evidence that fireworks can pollute both the air and water in ways that extend beyond the visible smoke。 The findings show that leftover debris, fine particles, and airborne chemicals may affect ecosystems and increase people's exposure to air pollution during major celebrations
The class of radio sources known as Compact Symmetric Objects (CSOs) is of particular interest in the study of the evolution of radio galaxies. CSOs are thought to be young (probably ~10^4 years), and a very high fraction of them exhibit HI absorption toward the central parsecs. The HI, which is thought to be part of a circumnuclear torus of accreting gas, can be observed using the VLBA with high enough angular resolution to map the velocity field of the gas. This velocity field provides new information on the accretion process in the central engines of these young sources. We have identified 9 new CSOs from radio continuum observations for the VLBA Calibrator Survey, increasing the number of known CSOs by almost 50%.