Wed 17 May 2023 16:52 - 17:00 at Meeting Room 104 - Development and evolution of AI-intensive systems Chair(s): Sebastian Elbaum
Compositional AI systems, which combine multiple artificial intelligence components together with other application components to solve a larger problem, have no known pattern of development and are often approached in a bespoke and ad hoc style. This makes development slower and harder to reuse for future applications. To support the full rapid development cycle of compositional AI applications, we have developed a novel framework called (Bee)∗ (written as a regular expression and pronounced as “beestar”). We illustrate how (Bee)∗ supports building integrated, scalable, and interactive compositional AI applications with a simplified developer experience.
Wed 17 MayDisplayed time zone: Hobart change
Wed 17 May
Displayed time zone: Hobart change
15:45 - 17:15 | Development and evolution of AI-intensive systemsSEIP - Software Engineering in Practice / Technical Track / NIER - New Ideas and Emerging Results at Meeting Room 104 Chair(s): Sebastian Elbaum University of Virginia | ||
15:45 15mTalk | Reusing Deep Neural Network Models through Model Re-engineering Technical Track Binhang Qi Beihang University, Hailong Sun Beihang University, Xiang Gao Beihang University, China, Hongyu Zhang The University of Newcastle, Zhaotian Li Beihang University, Xudong Liu Beihang University | ||
16:00 15mTalk | PyEvolve: Automating Frequent Code Changes in Python ML Systems Technical Track Malinda Dilhara University of Colorado Boulder, USA, Danny Dig JetBrains Research & University of Colorado Boulder, USA, Ameya Ketkar Uber Pre-print | ||
16:15 15mTalk | DeepArc: Modularizing Neural Networks for the Model Maintenance Technical Track xiaoning ren , Yun Lin Shanghai Jiao Tong University; National University of Singapore, Yinxing Xue University of Science and Technology of China, Ruofan Liu National University of Singapore, Jun Sun Singapore Management University, Zhiyong Feng Tianjin University, Jin Song Dong National University of Singapore | ||
16:30 15mTalk | Decomposing a Recurrent Neural Network into Modules for Enabling Reusability and Replacement Technical Track Sayem Mohammad Imtiaz Iowa State University, Fraol Batole Dept. of Computer Science, Iowa State University, Astha Singh Dept. of Computer Science, Iowa State University, Rangeet Pan IBM Research, Breno Dantas Cruz Dept. of Computer Science, Iowa State University, Hridesh Rajan Iowa State University Pre-print | ||
16:45 7mTalk | Safe-DS: A Domain Specific Language to Make Data Science Safe NIER - New Ideas and Emerging Results Pre-print | ||
16:52 7mTalk | Rapid Development of Compositional AI NIER - New Ideas and Emerging Results Lee Martie MIT-IBM Watson AI Lab, Jessie Rosenberg IBM, Veronique Demers MIT-IBM Watson AI Lab, Gaoyuan Zhang IBM, Onkar Bhardwaj MIT-IBM Watson AI Lab, John Henning IBM, Aditya Prasad IBM, Matt Stallone MIT-IBM Watson AI Lab, Ja Young Lee IBM, Lucy Yip IBM, Damilola Adesina IBM, Elahe Paikari IBM, Oscar Resendiz IBM, Sarah Shaw IBM, David Cox IBM Pre-print | ||
17:00 7mTalk | StreamAI: Challenges of Continual Learning Systems in Production for AI Industrialization SEIP - Software Engineering in Practice Mariam Barry BNP Paribas, Albert Bifet University of Waikato, Institut Polytechnique de Paris, Jean Luc Billy BNP Paribas | ||