Waterloo Research’s Post

Update: the position is not available anymore :) PhD Opportunity in AI for Legal Reasoning / Opportunité de doctorat en Intelligence Artificielle appliquée au raisonnement juridique We are searching for a candidate with a Master degree in Computer Science, with knowledge about Artificial Intelligence (Knowledge Representation and Reasoning, Machine Learning, Natural Language Processing) and a strong interest for interdisciplinary work. The PhD student will work on - the definition of argumentation formalisms allowing the representation of context-sensitive information (e.g. contextual preferences between arguments); - the extraction of such argumentative information from legal text. The PhD candidate will be part of the AIDAL research group (https://www.irit.fr/aidal/), a newly founded team in AI and Law at the Université Toulouse Capitole. The scholarship covers 3 years, with approximate salary 2100EUR (gross) , with the possibility of topping up with teaching activities. Ideal starting date in Autumn 2024. More information is available here: https://lnkd.in/gAy2ycYR #ArtificialIntelligence #AI #LegalReasoning #PhD #IntelligenceArtificielle #IA #RaisonnementJuridique #doctorat #thèse IRIT (Institut de Recherche en Informatique de Toulouse) Université Toulouse Capitole Laurent Perrussel

With the start of the fall term the second PhD program “Computational X” is launched at IT:U and impressively demonstrates the interdisciplinarity of the new Technical University in Linz (Upper Austria). ➡ The research fields of the eleven founding professors range from Neuroscience to Human Computing Interaction, from Natural Language Processing to Geosocial AI. ➡ The PhD program Computational X promotes interdisciplinary research facilitating the convergence of natural sciences like biology, social sciences and humanities with computational sciences. ➡ The program supports junior researchers in developing their abilities to perform independent scientific research. ➡ PhD students will acquire a comprehensive range of computational skills and methods, ranging from data science to artificial intelligence to robotics, as well as a solid core of 21st century skills. 💡 This will empower students to transform their respective disciplines with computational methods and explore frontiers in the digitalization of natural sciences, social sciences and humanities. The eleven founding professors are: ✅ Ass. Prof. Sebastian Dennerlein– Understanding and Supporting (inter-)professional Transformations ✅ Prof. Christopher Frauenberger – Designing Human-Computation Relationships ✅ Prof. Christian Hilbe - Game Theory and Evolutionary Dynamics ✅ Prof. Yufang Hou – Natural Language Processing (NLP) ✅ Ass. Prof Nina Christine Hubig- Explainable Artificial Intelligence ✅ Ass. Prof. Jie Mei– Computational Neuroscience ✅Ass. Prof. Daniel Klotz– Machine Learning in Earth Science ✅ Prof. Tiago Peixoto– Complex Systems and Network Science ✅ Prof. Bernd Resch - Geosocial Artificial Intelligence ✅ Prof. Ben Wagner – Human Rights and Technology ✅ Prof. Philipp Wintersberger – Intelligent User Interfaces This PhD program marks another significant step in IT: U’s mission to shape the future of higher education and scientific research. 💡 #PhDProgram #InterdisciplinaryResearch #DigitalTransformation #HigherEducation #AI #NLP #HCI #GameTheory #XAI #ComputationalNeuroscience #EnvironmentalAI #GeoAI #HumanAICollaboration #TransformFutures

Staying updated with the latest news is crucial as the AI industry continues to evolve. Here are three noteworthy developments that have recently emerged. First, an exciting update from the ACL Anthology, a digital archive of research papers in computational linguistics sponsored by the Association for Computational Linguistics (ACL). They have recently added missing long and short papers to their anthology. This enrichment not only broadens the repository but also significantly benefits researchers worldwide who rely on this comprehensive collection for their work. Such expansions are essential as they provide more resources for advancing research and innovation in computational linguistics. Next, there is an ongoing recruitment drive for science teachers led by Hari Subramonyam, a prominent figure in AI. The call for applications highlights the critical need for competent educators in science fields. As we advance further into the era of AI, having knowledgeable science teachers becomes even more important. These educators inspire future generations who will continue to push the boundaries of technology and innovation. Lastly, representatives from Mila, a Montreal-based institute dedicated to AI research, are set to participate in the DeepIndaba event in Dakar. DeepIndaba aims to strengthen machine learning and AI across Africa by fostering community among African practitioners. Mila's involvement underscores how international cooperation can propel industry advancements forward and highlights the global nature of progress in this field. These stories illustrate how expanding academic resources, recruiting skilled educators, and fostering international collaborations play crucial roles in shaping our future with AI technology.#AI #MachineLearning #Education

Congratulations to six CDS PhD students selected for the 2024-2025 Meta AI Mentorship Program: Angelica Chen, Guy Davidson, Jingtong Su, Nikolaos Tsilivis, Jianyu Zhang, and Lily Zhang. They join an accomplished group of CDS PhD students selected in previous years: 2023-2024: Polina Kirichenko, Aram-Alexandre Pooladian, Vlad Sobal, Jingtong Su, Jianyu Zhang, and Yanli Zhou. 2022-2023: Polina Kirichenko, Sanae Lotfi, Aram-Alexandre Pooladian, Vlad Sobal, and Yanli Zhou. The program pairs students with AI at Meta researchers for 12-24 months of collaborative work at Meta's New York office. Selected from a competitive pool of applicants, these students conduct research in computer vision, machine learning, natural language processing, and related fields. #MetaAI #DataScience #ArtificialIntelligence #MachineLearning #AIResearch

Summer Schools: Summer School in Language and Translation Technologies: The first edition of the Summer School on Language and Translation Technologies, aimed at students, PhD graduates, PhD candidates, professionals, teachers, and academics, offers a mix of theoretical lessons and practical workshops designed to explore translation processes supported by artificial intelligence. The program will cover neural network models and the evolution of machine translation, including post-editing stages and quality assessment of translations. Additionally, aspects related to

CDS Research Scientist Ravid Shwartz Ziv is pioneering a new model of scientific collaboration by coordinating multiple research projects through a public Discord server. This initiative, which began as a simple Twitter post, has grown into a community where anyone interested can contribute to exploring connections between large language models and information theory. The server currently hosts several active projects, including studies on neural network compression and the effects of intermediate layers in large language models. One of these projects has already resulted in a paper accepted at NeurIPS. This experimental model allows researchers from diverse backgrounds to collaborate freely, pushing the boundaries of traditional institutional research. 🔗 Read more: https://lnkd.in/d9ykv5F4

This fall, I will contribute one course day to our new Generative AI CAS. In case you're interested in learning how GAI works, how to properly, ethically and securely apply it in practice, and learn from leading researchers, we're organizing a webcast on May, 6 where you can learn more about the CAS.

Call for Papers: Special Issue on Application of Large Language Models in Materials Science We are excited to announce a special issue in the Materials Today Communications journal titled "Application of Large Language Models in Materials Science." Submission Deadline: 31 October 2024 Submit your paper: https://lnkd.in/e_UxEpFw Mehrdad Jalali and Enrico Salvati edited this special issue, which aims to highlight innovative applications of large language models (LLMs) in materials science. We invite researchers and practitioners to contribute their groundbreaking work to this issue. Your paper can provide insights into how LLMs can transform materials discovery, simulation, data analysis, and more. Topics of interest include but are not limited to: - LLMs in Materials Informatics - Natural Language Processing for Materials Research - Generative AI Applications in Materials Design and Characterization Join us in exploring the frontiers of AI in materials science. We look forward to receiving your submissions! #MaterialsScience #AI #LLMs #SpecialIssue #CallForPapers #MTCOMM #Elsevier

I'm thrilled to announce my latest research work in field of artificial intelligence In our study, we explore the ability of LLMs to generate novel research ideas across five key domains: Computer Science, Economics, Chemistry, Physics, and Medical Sciences. Paper link: https://lnkd.in/gMfg3aV5 Key highlights of our work: 1.  We developed a novel dataset of recent papers across multiple fields to benchmark the models. 2. Introduced new evaluation metrics – Idea Alignment Score and Idea Distinctness Index – to assess the quality of ideas generated. 3. Our findings show that Claude-2 and GPT-4 outperform GPT-3.5 and Gemini, with Claude-2 leading in generating more diverse future research ideas. This research highlights both the potential and limitations of AI in helping researchers develop future-oriented, innovative concepts. Special thanks to Sandeep Kumar who supported this work. Looking forward to further exploring the evolving role of AI in research!

I am happy to share that I received a First Class for my undergraduate research project titled: "Large Language Model Parameter Efficient Fine-Tuning for Mathematical Problem Solving." In this project, I explored the utilization of a parameter-efficient fine-tuning technique, QLoRA, to enhance the ability of large language models to solve mathematical problems. I compared QLoRA with both no fine-tuning and full fine-tuning. By focusing on parameter efficiency, my research aimed to reduce computational resource usage while maintaining high performance in problem-solving tasks. Key Finding: QLoRA achieved notable improvements in solving mathematical problems and outperformed traditional fine-tuning approaches, all with a fraction of the computational costs and time required for training large language models. This research has important implications for the future of AI in education and various industries where mathematical problem-solving is crucial. I'm excited about the potential applications and further developments in this field. A big thank you to my advisor, professors, and peers who supported me throughout this journey. Looking forward to contributing more to the field of AI and machine learning! #AI #MachineLearning #Research #Education #Mathematics #FineTuning #LanguageModels #ParameterEfficiency #LoRA #QLoRA