Large Language Models (LLMs) are fast becoming indispensable tools for software developers, assisting or even partnering with them in crafting complex programs. The advantages are evident -- LLMs can significantly reduce development time, generate well-organized and comprehensible code, and occasionally suggest innovative ideas that developers might not conceive on their own. However, despite their strengths, LLMs will often introduce significant errors and present incorrect code with persuasive confidence, potentially misleading developers into accepting flawed solutions. In order to bring LLMs into the software development cycle in a more reliable manner, we propose a methodology for combining them with ``traditional'' software engineering techniques in a structured way, with the goal of streamlining the development process, reducing errors, and enabling users to verify crucial program properties with increased confidence. Specifically, we focus on the Scenario-Based Programming (SBP) paradigm -- an event-driven, scenario-based approach for software engineering -- to allow human developers to pour their expert knowledge into the LLM, as well as to inspect and verify its outputs. To evaluate our methodology, we conducted a significant case study, and used it to design and implement the Connect4 game. By combining LLMs and SBP we were able to create a highly-capable agent, which could defeat various strong existing agents. Further, in some cases, we were able to formally verify the correctness of our agent. Finally, our experience reveals interesting insights regarding the ease-of-use of our proposed approach. The full code of our case-study will be made publicly available with the final version of this paper.

Deep neural networks (DNNs) have become a crucial instrument in the software development toolkit, due to their ability to efficiently solve complex problems. Nevertheless, DNNs are highly opaque, and can behave in an unexpected manner when they encounter unfamiliar input. One promising approach for addressing this challenge is by extending DNN-based systems with hand-crafted override rules, which override the DNN's output when certain conditions are met. Here, we advocate crafting such override rules using the well-studied scenario-based modeling paradigm, which produces rules that are simple, extensible, and powerful enough to ensure the safety of the DNN, while also rendering the system more translucent. We report on two extensive case studies, which demonstrate the feasibility of the approach; and through them, propose an extension to scenario-based modeling, which facilitates its integration with DNN components. We regard this work as a step towards creating safer and more reliable DNN-based systems and models.

An Israeli biotechnology innovation lab set up last year and backed by some of the world's leading pharmaceutical companies like Pfizer and Merck has formed a new startup that will harness artificial intelligence (AI) to assess drug efficacy in pre-clinical trials and improve chances for success in later stages. The startup, OMEC.AI, is the first company established with funding and support from AION Labs, a Rehovot-based organization launched last October with a mission to create and invest in early-stage startups focused on AI and computational biology in drug discovery and development. AION Labs is a collaboration between pharmaceutical giants Pfizer, AstraZeneca, Merck, and Teva Pharmaceuticals, together with Amazon's AWS and the Israel Biotech Fund, and is headed by Mati Gill, a former senior executive at Teva, and Dr. Yair Benita, the former head of computational biology at Compugen, science operations at CytoReason, and principal scientist at MSD (Merck). AION Labs ran three bootcamps over the past year to field scientist founders and inventors focused on addressing key industry challenges identified by the global pharma companies such as designing antibodies for targeted therapeutics and analyzing data using AI to assess and predict the clinical readiness of drug candidates. The latter challenge produced OMEC.AI, founded this summer with $2 million (NIS 7 million) in seed funding by AI experts Ori Shachar and Amir Harel, both of whom previously led AI teams at Mobileye, Intel's Jerusalem-based autonomous driving subsidiary.

The developed world is on the cusp of turning the corner in the fight against COVID-19 thanks to the unprecedented effort to rapidly develop and distribute effective vaccines. Now technologists are hoping to take drug development to the next level, and AI will play a big role. One of the companies at the forefront of using machine learning and AI to develop drugs is CytoReason. The company helps pharmaceutical firms like Pfizer accelerate drug development by providing high resolution models of the human body that's infected with the disease that the drug companies are targeting. "If I told you that in 200 years, drugs would be developed in a computer, you would not be real surprised," said CytoReason CEO and founder David Harel.

Cybersecurity is a major concern for every business across all verticals. Software vulnerabilities and targeted attacks are two major cybersecurity concerns for today's modern business. The latest solutions in artificial intelligence (AI) and machine learning (ML) are being implemented to aid in preventing cybersecurity attacks and securing software vulnerabilities. According to market research by Mordor Intelligence, the global cybersecurity market was valued at USD 161.07 billion in 2019, and is expected to reach USD 363.05 billion by 2025, registering a compound annual growth rate of 14.5 percent during the period of 2020 to 2025. The research company says the popularity of the internet of things, bring your own device, artificial intelligence (AI), and machine learning (ML) in cybersecurity is increasing, leading to more vulnerabilities and an ever-growing need to secure networks and devices.

Developing a new drug can cost billions of dollars and take a dozen or more years to bring to market. Two Israeli researchers have applied artificial intelligence (AI) and deep learning to shave time and money off the drug-discovery process. Instead of searching for the appropriate molecules to use in a new medicine, as is done today, they enabled a computer to make smart predictions without human guidance. Shahar Harel and Kira Radinsky at the Technion-Israel Institute of Technology fed into their computer system hundreds of thousands of known molecules as well as the chemical composition of all FDA-approved drugs up until 1950. Aided by AI, the computer came up with new potential molecules by making sometimes unexpected correlations from within this massive sample.

When it comes to human senses, we've found ways to reproduce the look and sound of the real world reasonably accurately. There are even technologies for reproducing the feel of certain experiences, such as flight and car simulators. But the problem of reproducing smell is much more intractable. The 1960 SmelloVision experiment is a case in point. This involved some 30 odors that were released into the cinema at certain times during a movie.

In the visual and aural realms, we very often interact with reproduced versions of an original -- a photograph of a scene, a recording of a concert. And as long as you know what the original looked and/or sounded like, it's easy to tell whether it's an accurate reproduction. For smells, the same does not hold true. Unlike audio or visual reproductions, it's hard to transmit a reproduction of a smell to someone. There have been a few attempts, of course, but while something visual can be mimicked using wavelength and luminance and sound is a matter of copying the tone, odors depend on the brain's perception of molecules.