213476 research outputs found
Sort by
SurgRAW: Multi-Agent Workflow with Chain of Thought Reasoning for Robotic Surgical Video Analysis
No full textRobotic-assisted surgery (RAS) is central to modern surgery, driving the need for intelligent systems with accurate scene understanding. Most existing surgical AI methods rely on isolated, task-specific models, leading to fragmented pipelines with limited interpretability and no unified understanding of RAS scene. Vision-Language Models (VLMs) offer strong zero-shot reasoning, but struggle with hallucinations, domain gaps and weak task-interdependency modeling. To address the lack of unified data for RAS scene understanding, we introduce SurgCoTBench, the first reasoning-focused benchmark in RAS, covering 14256 QA pairs with frame-level annotations across five major surgical tasks. Building on SurgCoTBench, we propose SurgRAW, a clinically aligned Chain-of-Thought (CoT) driven agentic workflow for zero-shot multi-task reasoning in surgery. SurgRAW employs a hierarchical reasoning workflow where an orchestrator divides surgical scene understanding into two reasoning streams and directs specialized agents to generate task-level reasoning, while higher-level agents capture workflow interdependencies or ground output clinically. Specifically, we propose a panel discussion mechanism to ensure task-specific agents collaborate synergistically and leverage on task interdependencies. Similarly, we incorporate a retrieval-augmented generation module to enrich agents with surgical knowledge and alleviate domain gaps in general VLMs. We design task-specific CoT prompts grounded in surgical domain to ensure clinically aligned reasoning, reduce hallucinations and enhance interpretability. Extensive experiments show that SurgRAW surpasses mainstream VLMs and agentic systems and outperforms a supervised model by 14.61% accuracy. Dataset and code is available at https://github.com/jinlab-imvr/SurgRAW.git
Harnessing Controlled Dealloying-Support Coupling for Ultrastable PtNi Catalysts in PEMFC Applications
No full textPlatinum–transition metal (PtM) alloys are among the most promising oxygen reduction reaction (ORR) catalysts, yet their practical deployment in proton-exchange membrane fuel cells (PEMFCs) is hindered by transition-metal dissolution, particle coarsening, and insufficient durability. Moreover, conventional alloying or intermetallic ordering strategies often aggravate these issues by inducing severe nanoparticle aggregation and instability. Here we report a controllable alloying–dealloying strategy to construct PtNi nanoparticles confined in an N-doped carbon framework (Pt1Ni1-x@Nix_NC). Ammonia-assisted dealloying produces a Pt-rich shell with an alloyed core, while the N-doped carbon anchors the released Ni atoms form Ni–N/C moieties, thereby suppressing agglomeration and strengthening metal–support interactions. This coordination–support coupling optimizes Pt 5d orbital occupation, weakens oxygen adsorption, and accelerates ORR kinetics. Consequently, Pt1Ni1-x@Nix_NC exhibits a half-wave potential of 0.932 V and an ultrahigh mass activity of 2.028 A mgPt−1, which is 8.75-fold higher than commercial Pt/C and among the best values reported to date for PtNi-based catalysts. Remarkably, it shows only a 6 mV half-wave potential loss after 30,000 cycles, demonstrating exceptional durability. In PEMFCs, the fuel cell delivers 975 mW cm−2 peak power density and retains 91.9% of initial performance, underscoring a generalizable approach for designing durable, high-performance low-PGM catalysts for next generation PEMFCs
Evidence for the Keplerian orbit of a close companion around a giant star
No full textClose companions influence stellar evolution through tidal interactions, mass transfer and mass-loss effects. While such companions are detected around young stellar objects, main-sequence stars, red giants and compact objects, direct observational evidence of close-in companions around asymptotic giant branch (AGB) stars has remained elusive. Here we present (sub)millimetre time-domain imaging spectroscopy revealing the Keplerian motion of a close-in companion around the AGB star π1 Gruis. The companion, slightly more massive than the AGB star, is likely a main-sequence star. Unlike more evolved stars with companions at comparable distances, the companion of π1 Gruis follows a circular orbit, suggesting an eccentricity-generating mechanism during the late- or post-AGB phase. Our analysis suggests that model-predicted circularization rates may be underestimated. Our results highlight the potential of multi-epoch (sub)millimetre interferometry in detecting the Keplerian motion of close companions to giant stars and open avenues for our understanding of tidal interaction physics and binary evolution
Comparative analysis of microRNA expression in serum and plasma in patients screened for BRCA1 or BRCA2 mutations
No full textMicroRNA-based biomarkers hold promise for precision oncology, but standardization and pre-analytical variability remain barriers to clinical use. We performed a technical study to assess how the type of biological material (plasma, serum) and quantification (miRNA-seq, RT-qPCR) method affect microRNA (miRNA) levels. Samples from 202 women – 154 carriers of BRCA1/2 pathogenic variants, from three centers were collected. MiRNAs were quantified using miRNA sequencing (miRNA-seq), a quantitative PCR (qPCR) panel of 182 miRNAs (Qiagen) or targeted qPCR of 13 miRNA associated with BRCA mutations or ovarian cancer. Differential expression and linear regression analyses were used to assess methods concordance and biological material-related differences. Differences of individual miRNA levels between serum and plasma were more pronounced in sequencing than in qPCR. miR-193a-3p, miR-197-3p, miR-877-5p and miR-2110 were detected in significantly higher amounts in serum, while miR-28-5p in plasma, by both sequencing and qPCR. Measurements done by sequencing and qPCR correlated positively and significantly for ~ 75% of investigated miRNA. Finally, strong inter-assay correlation of established qPCR hemolysis marker (based on miR-451a and miR-23a-3p) provided evidence that it can be used also in sequencing data. Differences in absolute and relative miRNA levels in serum and plasma necessitate material-specific test calibration and may impact selection of best miRNA biomarkers for specific scenarios. Concordance in expression levels quantified by miRNA-seq and qPCR is generally good for high-expression miRNAs in both serum and plasma, but a conversion of multi-miRNA tests from serum to plasma is unfeasible without experimental calibration
Tottenham / Durham Learning Exchange: Building understanding, solidarity and cooperation across difference
No full textThis pamphlet showcases the findings and proposals emerging from a unique learning exchange between community organisations in Durham and Tottenham. In both Tottenham and Durham, community groups are advancing their own plans to repurpose and manage social infrastructure to support place- based economic and social development that meets local needs and desires. While they do so in different contexts, the Learning Exchange demonstrates that they share common purpose, values and challenges. This pamphlet calls for a profound strengthening of place-based grassroots alliances alongside inter-place networks and coalitions to facilitate understanding and solidarity across difference, impacting regional development policy debates. It makes seven calls to action which we hope will be taken up by communities, policy-makers, universities and funders in various ways. This is a continued piece of work from UCL's Pro-Provost Regional Communities (https://www.ucl.ac.uk/about/leadership/organisation/president-provosts-team/pro-provost-regional-communities)
Beacons and Military Communication from Antiquity to the Early Modern Period
No full textFrom fire-lit beacons to sweeping lines of sight, this book uncovers the hidden infrastructure of pre-modern warfare. It is the first archaeological study to trace how military signalling systems shaped landscapes and societies from antiquity to the early modern period. Spanning ancient Greece, Viking Age Scandinavia, and medieval North Africa, the book investigates how visual communication enabled mobilisation, defence, and control. Through case studies and new interpretations of archaeological and historical data, it reveals a world where ordinary people, not just armies, played a central role in responding to the threat of war. A major contribution to military and landscape archaeology.
Contributors are Jørgen Bakke, Alexandre Ausonius Bertaud, Stuart Brookes, Tiffany Earley-Spadoni, Martin Malcolm Elbl, Andreu Galera Pedrosa, Frode Iversen†, Andrew Johnson, Christian Juel, Thorsten Lemm, Lucas R. McMahon, Mette Stauersbøl Mogensen, Marie Ødegaard, Ainhoa Pancorbo Picó, Mads Ravn, Mads Runge, Andreas Thiel, and Joey L. Williams
Technical Report on the Design, Organisation and Evaluation of the Second Citizens' Assembly on Energy in Hamra
No full textThe Citizens’ Assembly (CA) on electricity and energy justice in Hamra was conducted
over three days in November 2025 in Hamra, Beirut. It was organised and implemented
by Jibal NGO, following a partnership with and methodological input from the
(En)Visioning Justice project at University College London. It builds on the first CA held
on energy justice in Hamra in 2020and serves as a second iteration aimed at refining the
methodology within a real deliberative setting and contributes to the creation of a
Toolkit for Organising Citizens’ Assemblies in Arab Countries. In 2020, Lebanon was
witnessing a deteriorating decrease of public electricity provision, generator providers
were progressively stepping in to fill the gap. A year later, the government removed fuel
subsidies, leading to diesel shortages, long power cuts and rapidly rising fuel prices.
Households resorted to either buying their own private generators or to installing their
own solar systems to meet their electricity needs; while many people were losing
access to electricity. With air pollution on the rise, people were organizing in different
ways, and there was a call to create a collective context that is fair for everyone. The
second CA’s primary purpose was then to understand the current energy context in
Lebanon and to expose the residents of Hamra to existing and emerging alternative
energy models; and through structured deliberation and democratic participation, to
identify collective priorities and criteria for energy justice in Hamra, ultimately reaching
final recommendations through a voting process.
The CA centered on one overarching question: How can we organise collectively to reduce the cost, pollution, and mental burden of electricity on our lives? To answer this,
35 randomly selected residents of Hamra participated in a structured process
combining learning, facilitated discussions, and decision-making. Jibal conducted
comprehensive planning: consultations with experts, local stakeholders, and
community actors; the formation of an advisory group and evidence group; a
randomized recruitment process; and a three-day convening during which assembly
members (participants) deliberated and developed recommendations. Through this
process, participants identified key criteria to be considered and recommendations for
more equitable and sustainable energy provision. These recommendations are
intended to inform Beirut’s wider energy stakeholders discussions through the
assembly’s observers; activists, municipal representatives, government institutions,
and private sector actors; pushing for the inclusion of perspectives which are often
excluded from formal decision-making to be meaningfully integrated into future energy
planning and policy making
Quantifying forest structural attributes and aboveground carbon dynamics with terrestrial laser scanning in a temperate deciduous forest
No full textQuantifying forest structure and aboveground biomass carbon (AGBC) dynamics over time is crucial for evaluating climate change impact on carbon stocks, and providing key insights into changes in the terrestrial carbon cycle. To date, the use of multi-temporal terrestrial laser scanning (TLS) to detect temporal dynamics of forest structure and AGBC remains largely unexplored. In this study, we demonstrate the use of bi-temporal TLS data to quantify fine-scale dynamics of forest structure and AGBC. A total of 831 live trees were extracted and manually aligned from two leaf-off datasets collected in a 1.4 ha area of temperate woodland (Wytham Woods, UK) in 2016 and 2022. Results indicated that, at the individual tree level, most trees exhibited positive growth in structural attributes between 2016 and 2022, including diameter at breast height (DBH, 60.2 % of trees), tree height (H, 75.8 %), crown projection area (CPA, 64.7 %), crown volume (CV, 60.5 %), and aboveground volume (V, 50.5 %). At the plot level, all structural attributes also increased, including basal area (1.8 m²/ha, 4.8 % growth), H (128.9 m/ha, 1.4 %), CPA (411.9 m²/ha, 3.0 %), DBH (1.5 m/ha, 1.1 %), CV (181.7 m³/ha, 0.3 %), and V (7.9 m³/ha, 1.0 %). The total AGBC of the study area saw a net carbon gain of 0.4 Mg C/ha/year over the six-year period. Notably, trees with DBH greater than 60 cm contributed over 40 % of the total AGBC. Moreover, our results reveal that branch dynamics play a crucial role in AGBC dynamics, underscoring the added value of TLS for tracking AGBC changes over time
For everyday arguments prior beliefs play a larger role on perceived argument quality than argument quality itself
No full textNot all arguments are equally convincing, and whilst a given argument may be persuasive to some people, it is often seen as inadequate by others. We are interested in both the individual and argument level differences that make ‘everyday’ arguments such as those on social media persuasive. We investigate this question using a paradigm that consists of two parts. In the first part, we measure participants' individual beliefs about eight claims each referring to a political topic (e.g., Abortion should be legal). In the second part, participants rated the quality of an argument for each of these claims. Arguments were good or bad (Experiments 1 and 2) or good, inconsistent, or authority-based (Experiment 3). Good, inconsistent, and authority-based arguments summarised arguments from an educational bipartisan website, contained internal inconsistencies, or were based on appeals to authority, respectively. We found that participants preferred arguments that were also in line with their beliefs. We also found that participants were able to discriminate the qualities of different arguments – good arguments were rated as better than any other type of argument. In Experiment 3, inconsistent arguments were rated as better than those making appeals to authority. Importantly, the maximum effect of belief was larger than the maximum effect of argument quality. Thus, people do not evaluate arguments independently of the background beliefs held about them, which play at least as large a role in evaluating the quality of the argument as does the actual quality of the argument itself
Sweet music: creative collaboration with laryngectomy patients
No full textThis research blog explores how Professor Evangelos Himonides (UCL Institute of Education) and Dr Thomas Moors, founder of the charity Shout at Cancer, have spent over a decade working alongside people who have undergone total laryngectomy to transform rehabilitation through music and creative practice. Rather than treating patients solely as research participants, their work represents genuine co-creation, bringing together researchers, clinicians, artists, and people with lived experience to experiment with singing, beatboxing, performance, visual art and even food design. The collaboration not only supports communication, confidence, and wellbeing after life-changing surgery, but also advances understanding of voice, technology, and inclusive research methods. The blog highlights how public engagement, patient and public involvement, and co-production can intersect to produce research that is academically robust, socially meaningful, and creatively ambitious