Our new preprint is out: we use CatBoost and Graph Attention Networks (GATs) to accurately predict holding maneuvers while offering interpretability through graph-based feature importance.
https://t.co/qji1Xahdn4
In a new paper, we use network analysis to examine community disputes in post-civil war Mozambique. By analyzing a range of conflicts, from domestic disputes to wartime accusations, we reveal the deep-rooted tensions that persist in the region.
https://t.co/qwd1u3B0Rw
Predicting a research paper's citation impact is crucial for stakeholders to assess the initial influence of a study and allocate resources to promising research. In our recent work, we compared the performance of GPT and other embedding models
https://t.co/3hrG22amlY
Today in @Nature, I laid out our vision of using AI to help universities maximize their research impacts.
Key Q: Can we use data and AI tools to help a university locate key ideas and personnel to maximize its impact and accelerate progress?
https://t.co/s8ITQsC9jS
🧵 1/
Our new paper is out: "Interconnectivity disrupted by fading globalization: A network approach to recent international trade developments."
https://t.co/kZ3nk0tvvb
Esta semana, a @IUCN publicou nosso relatório sobre o futuro dos óleos vegetais, os quais são muito mais importantes na nossa vida do que nós pensamos.
Veja mais informações neste texto, em português: https://t.co/C1urujIL4r
🧵1/4
In our new preprint on @arxiv , we show that including socio-economic data in epidemic forecasting can result in more accurate results. We consider data from Dengue, Zika, COVID-19 and Influenza. With @luizalober and Kirstin Roster.
https://t.co/SZW4PVty99
Have you ever wondered what distinguishes a fiction book that achieves bestseller status? We explored this question in our latest paper:
Using full-text content to characterize and identify best seller books: a study of early 20th-century literature
https://t.co/L0y9Qv2BSQ
In our recent paper, we uncover the evolution of topics and themes in Chemistry of Materials. Our latest editorial revisits the rich history of the journal, highlighting the most relevant subjects across three decades.
https://t.co/IBngYhJdr3
Our approach offers a fresh perspective on the history of Chemistry of Materials, providing valuable insights for researchers . We use network analysis and natural language processing, leveraging the power of AI to analyze citation networks and identify communities.
In our latest paper, we investigate the impact of limited information on recovering network properties. We analyzed sequences obtained from different types of random walk biases.
https://t.co/5Y2hKdLhO7
In our latest paper, we present a method to measure the extent to which modular documents conform to hierarchical templates. Paper w/ @LdaFCosta , A. Benatti and A C M. Brito.
https://t.co/aqciPy5R1R
#networkScience#textmining#graphs