2. CLANN

Overview

CLANN is a proposed LSH-based algorithm for solving the Approximate Nearest Neighbour problem with theoretical guarantees, based on the previous work done on PUFFINN. The algorithm is divided in two steps:

• Index Creation, the dataset is partitioned in clusters and in each cluster a PUFFINN index is built
• Query Algorithm, for each cluster the query algorithm runs the PUFFINN query algorithm and aggregates the results, stopping the search whenever it does serve no purpose
  In theory the approach should cut the hit-rate distribution of the LSH function, improving the overall result.

tab: Components
```dataview
table created AS "Created", summary AS "Summary"
from "PARA/AREAS/CLANN"
where contains(type, "area_note")
sort created DESC
```
tab: Projects
```dataview
table type AS "Type", Status AS "Status", Priority_Level AS "Priority_Level"
from "PARA/PROJECTS"
where contains(connections, this.file.link)
where type = "project_family" OR type = "project_note"
sort Status ASC
```
tab: Papers
```dataview
table status AS "Status"
	from "ZETA/LITERATURE"
where contains(connections, this.file.link)
sort type ASC
```
tab: Other
```dataview
table type AS "Type"
from "PARA/RESOURCES/DOCUMENTATIONS" OR "PARA/WORKSTATION"
where contains(connections, this.file.link)
where type = "documentation_note" OR type = "workstation_note"
sort type ASC
```

tab: Scheduled Meetings
```dataview
TABLE scheduled_date as "Scheduled Date", start_time as "Start Time", summary as "Summary"
from #area/clann
where contains(type,"meeting")
sort meeting_status asc, scheduled_date asc
```

tab: Ongoing Task
```tasks
not done
tags include #area/clann
path does not include "SYSTEM"
sort by due date
```

tab: Completed Tasks
```tasks
done
tags include #area/clann
path does not include "SYSTEM"
sort by due date
```