Benjamin Harack

ORCID: 0000-0002-1813-7772
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Graphene research and applications
  • Topic Modeling
  • Digital Platforms and Economics
  • Machine Learning in Healthcare
  • Artificial Intelligence in Healthcare and Education
  • Low-power high-performance VLSI design
  • Blockchain Technology Applications and Security
  • Molecular Junctions and Nanostructures
  • Quantum and electron transport phenomena
  • Ethics and Social Impacts of AI
  • Artificial Intelligence in Healthcare
  • Advanced Memory and Neural Computing

Human Immunome Project
2024

The Human Diagnosis Project
2024

University of Oxford
2024

McGill University
2012

This review examines the properties of graphene from an experimental perspective. The intent is to most important results at a level detail appropriate for new graduate students who are interested in general overview fascinating graphene. While some introductory theoretical concepts provided, including discussion electronic band structure and phonon dispersion, main emphasis on describing relevant experiments as well novel applications In particular, this covers synthesis characterization,...

10.5402/2012/501686 article EN ISRN Condensed Matter Physics 2012-04-26

This review examines the properties of graphene from an experimental perspective. The intent is to most important results at a level detail appropriate for new graduate students who are interested in general overview fascinating graphene. While some introductory theoretical concepts provided, including discussion electronic band structure and phonon dispersion, main emphasis on describing relevant experiments as well novel applications In particular, this covers synthesis characterization,...

10.48550/arxiv.1110.6557 preprint EN other-oa arXiv (Cornell University) 2011-01-01

Although large language models (LLMs), such as OpenAI GPT-4 or Google PaLM 2, are proposed viable diagnostic support tools even spoken of replacements for "curbside consults," past studies show that they may lack sufficient accuracy real-life applications. In an effort to improve their and reduce the risk misdiagnoses, we applied methods from field collective intelligence produce synthetic differential diagnoses aggregate answers individual commercial LLMs (OpenAI GPT-4, Cohere Command, Meta...

10.1056/aics2400502 article EN NEJM AI 2024-10-17

Background: Large language models (LLMs) such as OpenAI's GPT-4 or Google's PaLM 2 are proposed viable diagnostic support tools even spoken of replacements for "curbside consults". However, LLMs specifically trained on medical topics may lack sufficient accuracy real-life applications. Methods: Using collective intelligence methods and a dataset 200 clinical vignettes cases, we assessed compared the differential diagnoses obtained by asking individual commercial (OpenAI GPT-4, Google 2,...

10.48550/arxiv.2402.08806 preprint EN arXiv (Cornell University) 2024-02-13

Artificial intelligence systems, particularly large language models (LLMs), are increasingly being employed in high-stakes decisions that impact both individuals and society at large, often without adequate safeguards to ensure safety, quality, equity. Yet LLMs hallucinate, lack common sense, biased - shortcomings may reflect LLMs' inherent limitations thus not be remedied by more sophisticated architectures, data, or human feedback. Relying solely on for complex, is therefore problematic....

10.48550/arxiv.2406.14981 preprint EN arXiv (Cornell University) 2024-06-21

Society could soon see transformative artificial intelligence (TAI). Models of competition for TAI show firms face strong competitive pressure to deploy systems before they are safe. This paper explores a proposed solution this problem, Windfall Clause, where developers commit donating significant portion any eventual extremely large profits good causes. However, key challenge Clause is that must have reason join one. Firms also believe these commitments credible. We extend model with how...

10.48550/arxiv.2108.09404 preprint EN other-oa arXiv (Cornell University) 2021-01-01
Coming Soon ...