@CompHCILab

We are the Computational HCI Lab @UMichCSE. We study methods, applications, and ethics of Computational Modeling and AI in HCI.

 Ann Arbor, MI
Joined August 2018
  • Tweets62
  • Following23
  • Followers385
At #Ubicomp2023, @AntarAnindya presented a computational behavior model that forecasts end-of-day physical functioning of people with Multiple Sclerosis (MS) to support timely behavior-based clinical interventions. Work with Anna Kratz and @nikola_banovic. doi.org/10.1145/3580887
Anindya presenting in front of a lectern. A sign reading UBICOMP / ISWC is behind the presenter, and the title slide is projected on a large screen in the right corner of the image.

ALT Anindya presenting in front of a lectern. A sign reading UBICOMP / ISWC is behind the presenter, and the title slide is projected on a large screen in the right corner of the image.

5
15
3,325
Former lab member @Xincheng_Huang presented StructureSense (a tracking system using UHF-RFID sensing that infers constructive assembly structures from object motion) at #Ubicomp2023. Work with Keylonnie Miller, Alanson Sample, and @nikola_banovic. doi.org/10.1145/3570343
Xincheng presenting in front of a lectern. A sign reading UBICOMP / ISWC is behind the presenter, and the title slide is projected on a large screen in the right corner of the image.

ALT Xincheng presenting in front of a lectern. A sign reading UBICOMP / ISWC is behind the presenter, and the title slide is projected on a large screen in the right corner of the image.

4
8
1,957
Congratulations to @nelescher for winning the Equity-centered Engineering Award at the 2023 Michigan Engineering Three Minute Thesis (3MT) Competition! (egs.engin.umich.edu/) Award sponsored by OCCE (culture.engin.umich.edu/).
Nel holding an enlarged blue check showing the University of Michigan Engineering logo, the name of the award (Equity-centered Engineering), and a three thousand dollars amount.

ALT Nel holding an enlarged blue check showing the University of Michigan Engineering logo, the name of the award (Equity-centered Engineering), and a three thousand dollars amount.

1
3
16
5,075
CompHCI lab at #CHI2023 in one place taking a break after @im__jane presented her paper. Left to right: @RuiyiWang153, @im__jane, @nikola_banovic, and @HossainTahera.
Four people standing in front of the Congress Center Hamburg. "Welcome to CHI23" sign is in the background and above their heads.

ALT Four people standing in front of the Congress Center Hamburg. "Welcome to CHI23" sign is in the background and above their heads.

4
1
35
4,095
Representing CompHCI Lab at #CHI2023 from left to right @HossainTahera, @nikola_banovic, and @im__jane.
Three people taking a selfie while holding their CHI 2023 badges.

ALT Three people taking a selfie while holding their CHI 2023 badges.

2
25
1,850
Congratulations to lab members @ananya_kasi and Zhe Chen who presented their undergraduate research posters at the Explore Computer Science Research (#ExploreCSR) event @UMichCSE! 1/n
Ananya Kasi stands in front of her poster showing her research and talks to Rada Mihalcea.

ALT Ananya Kasi stands in front of her poster showing her research and talks to Rada Mihalcea.

Zhe Che talks to Nikola Banovic in front of her poster showing Zhe's research.

ALT Zhe Che talks to Nikola Banovic in front of her poster showing Zhe's research.

7
9
39
3,170
"StructureSense: Inferring Constructive Assembly Structures from User Behaviors" uses Bayesian inference to detect what the user is building based their interaction with different parts of the structure. Led by now former lab members @Xincheng_Huang and Keylonnie Miller. #IMWUT
The user starts an assembly task wanting to building a wooden lamp. The parts are laid out on a table in front of the user. StructureSense tracks the structure and provides the next step instructions according to the user’s progress. The interface is mirrored to a tablet in the bottom-right corner and shows instructions for the first row of a multi tiered lamp.

ALT The user starts an assembly task wanting to building a wooden lamp. The parts are laid out on a table in front of the user. StructureSense tracks the structure and provides the next step instructions according to the user’s progress. The interface is mirrored to a tablet in the bottom-right corner and shows instructions for the first row of a multi tiered lamp.

An interface showing what happens when StructureSense detects that the user has made an error. In this example, the user has used the wrong (top) shelf instead of the correct (middle) shelf and is alerting the user about the error. The interface shows what StructureSense detected on the left and what the user should have constructed instead on the right.

ALT An interface showing what happens when StructureSense detects that the user has made an error. In this example, the user has used the wrong (top) shelf instead of the correct (middle) shelf and is alerting the user about the error. The interface shows what StructureSense detected on the left and what the user should have constructed instead on the right.

This image shows what happens after the user fixes the error from the previous step and proceeds with assembly. The image shows the user starting to build the next tier of the lamp by adding a wooden poll to it. The interface in the bottom corner shows the instructions for the current assembly step: adding four wooden polls to the current lamp tier.

ALT This image shows what happens after the user fixes the error from the previous step and proceeds with assembly. The image shows the user starting to build the next tier of the lamp by adding a wooden poll to it. The interface in the bottom corner shows the instructions for the current assembly step: adding four wooden polls to the current lamp tier.

The final structure showing fully assembled four-tier heigh wooden floor lamp.

ALT The final structure showing fully assembled four-tier heigh wooden floor lamp.

1
1
2
15
See a recent @UMichCSE interview with our lab members (in order of appearance: @nikola_banovic, @sumit_ast, @AntarAnindya, and @snehalbp) about our work on #ExplainableAI (XAI) (more specifically how to achieve #XAI through human-AI interaction): youtu.be/tGicpp08r20
8
2
21
"Automatically Labeling Low Quality Content on Wikipedia by Leveraging Patterns in Editing Behavior" improves detection of sentences that need citations, NPOV, and clarifications. Work by @sumit_ast, Sabrina Tobar Thommel, @halfak, and @nikola_banovic was accepted to #CSCW2021.
2
4
1
28
Load more