Group 25
Members: Bianca Cordazzo, Tianlei He, Xincheng Tan, Yilan Wang
For full documentation, instructions on how to install and how to use SmartDiff, please refer to docs/documentation.ipynb
Although great improvements have been made in the world of STEM with regards to diversity and inclusivity, it is no secret that a lot of work remains to be done to bridge the gap between technological advancements and underrepresented groups. As we designed this software, we brainstormed ways to make it as inclusive as possible, recognizing that people from diverse technical backgrounds might engage with it in the future. Therefore, we decided to implement a graphic user interface (GUI) to make the user experience as friendly and smoothly as possible. That being said, although interacting with the software is straightforward, the interface is designed for English speakers and assumes familiarity with elementary and multivariable functions, which could still act as barriers to future users. More work needs to be done to mitigate these effects as much as possible, and to make the user experience as informative and clear as possible.
Our software has both positive and negative implications, but here we will argue that the positive ones outnumber the negative ones. On a technical level, we worked on making this software as time- and memory- efficient as possible while also supporting complex operations involving higher order derivatives and multivariable functions. Since efficiency and scalability are important components of automatic differentiation, we believe we are contributing a new software to the broader community, making it more diverse. Also, as mentioned before, we directed our efforts to software inclusivity, though we recognize that more work in this area needs to be done. Finally, we are aware that job loss is a main concern that comes with several types of automation. However, instead of disrupting professional opportunities, we sincerely hope our tool will instead create opportunities for more people interested in using AD in their areas of research and work and are currently not doing due to a lack of user-friendly software. However, in the case of such disruption ever existing, efforts should be made to instruct people at risk on the usage of this software while making the transition as easy as possible.