ABCDEFGHIJKLMNOPQRSTUVWXYZAA
1
NameDate of event / ML+CoffeeWhen will you be attending?Problem descriptionDemo description
2
Amanda DeWitt4/10/20249-11amAt the Center for Health Disparities Research, our primary research study is The Neighborhoods Study. In this study, we collect historical address data for participants to investigate exposures over the life course. This data is generated in one of two ways – for decedent participants, we have a team of staff perform archival and public records abstraction to generate address timelines; for living participants, we collect this data via mail survey. Each method of data collection presents challenges in processing and preparing for analysis; mail survey respondents will often repeat similar information across pages or leave out critical information, and collecting data from records abstraction is costly and time consuming.

I often think that ML approaches could be of great use to our project – to name just a couple:
1. Entity Resolution for variants of the same entity
e.g. is Address X the same as Address Y, just spelled slightly differently, or highly abbreviated) is the John Doe listed in this source material the same John Doe we are looking for?

2. Constructing accurate event sequences based on available information
E.g. when did they move to address X? if they moved back to address X after living at address Y, can this type of complexity be handled accurately? What about seasonal addresses? What about when the respondent writes “same” or “til present”?

3. Can we leverage our existing GIS resources (ArcGIS Pro, curated historical geospatial data sources) in combination with ML approaches?

I would be extremely interested in discussing further at this event or others. One thing to note is that our data is PHI – but I could synthesize data for demonstration or discussion.
3
Linqi Lu and Xianshi Yu3/13/20249:30-10:30amXianshi Yu (MS student from CS) and I (PhD student from School of Journalism and Mass Communication) have been working on a project that involves using ML to analyze images and text data from social media in order to understand triggers of recreational cannabis products. We are looking forward to discussing our ML approach thus far with the community, and hope to gather some new ideas for future directions of work.
4
Will de la Bretonne2/14/20249-11amI work in a lab focused on improving crop varieties. We currently collect almost all of our traits by visual rating some examples of which include flowering time, plant height, leaf angle, bio mass, etc (the list goes on and on). I am interested in using computer vision with data collected from drones to automate as many of these trait collections as possible. If you have any insight into tools or methods that would be beneficial for this work I would appreciate your feed back. Keep on keeping on - Will
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100