The recently accepted InnoGuard (EU) and SwarmOps (SNF) projects will welcome 4 PhD students in Software Engineering & Robotics. Apply to the positions by sending the documents indicated here https://t.co/4QDYemLfyC to sebastiano.panichella@unibe.ch timo.kehrer@unibe.ch

Thrilled to announce the return of the UAV Testing Competition! 🚁 Join us at @ICSTconf and @sbftworkshop. Learn more: https://t.co/tf4Yfzwurw GitHub: https://t.co/d8HWVO8rCo Organized by: @SajadKhatiri, @Prasun_Saurabh, @zohditara, @dmytro_humeniuk, @spanichella.  

Awesome time at the workshop's social event. Such a cool opportunity to connect with everyone after a day of learning. Let's see what the second day of the workshop brings!

@zohditara talking about the DeepHyperion-UAV testing tool participating at the CPS-UAV testing tool competition at @sbftworkshop #UAV #Testing @ICSEconf

📊 I'm going to share our empirical study on misbehavior explanations for #DeepLearning at @ESEM_conf. It's statistically significant that you'll be there at 11:00 (Central Time) in Room Rhythms 3. Access the presentation slides here: https://t.co/M33HbWTcIP #ESEM2023

I am going to present our focused test generator for #DeepLearning systems at @issta_conf. See you at 11:15 (Pacific Time), Room G10. You can find the slides here: https://t.co/HiFALIdbyO

Different techniques for explaining misbehaviours of #DeepLearning based software produce dissimilar explanations. Our empirical comparison of low- and high-level explanations will be presented at @ESEM_conf. Co-authored by yt, @zohditara & @paolo_tonella.

In case you missed my keynote at ICSE 2023 entitled "The Road Toward Dependable AI Based Systems" and are interested in the topic, here are the slides:

☀ Hello #ICSE people! If you are looking for a beautiful summer school to send your students in September, check out #SIESTA2023 "Deus vs. Machina" at https://t.co/U29jFgjeYC 😎 @ICSEconf @icpcconf @msrconf #ICPC23 #ICPC2023 #MSR23 #MSR2023 #ICSE23 #ICSE2023

DL-based software testers may need to acquire new data with specific features observed in operation. DeepAtash, our latest test generator, can produce such inputs and will be presented at @issta_conf. Co-authored by yt, @zohditara and @paolo_tonella.

When and Why Test Generators for #DeepLearning Produce Invalid Inputs? Our empirical study involving both automated and human validators was just accepted at @ICSEConf. Co-authored by @paolo_tonella. A result of https://t.co/JQKb1dKWNd. @usisoftware

I am honoured to be among the #ERC10kGrantees. The @ERC_Research program gave me the opportunity to build the Precrime team. We are developing tools to quantify the uncertainty of AI components, predict and prevent misbehaviours, generate test cases. @USI_INF @usisoftware

Do you want to understand why your #deeplearning system misbehaves? Our #testing approach to explore and represent the feature space of DL systems will be presented at @issta_conf. Co-autored by @zohditara, Alessio Gambi and @paolo_tonella. @usisoftware @USI_INF