Welcome to the CS-E9410 Foundations of Pattern Discovery course in autumn 2026!
Everything about how the course runs, the study material, the tasks and the announcements that matter will be posted on this page. The slides of the opening lecture go deeper into prerequisites and day-to-day practicalities.
Overview
The course surveys the ideas behind finding structure in large collections of data and how far they carry on messy problems. It covers the main families of patterns and the algorithms that search for them: itemsets and rules, subgraphs, ways of grouping high-dimensional observations, mining of interaction networks, and the statistics that separate a genuine finding from an artefact of the search.
Lectures
Twelve lectures are held in Hall A1, always from 16:15–18:00. The slot is Tuesday for all but one of them — week 37 adds a Wednesday meeting on top of its Tuesday one, so read the table before turning up!
| Weekday | Date | Topic |
|---|---|---|
| Tuesday | 1.9. | L1: Introduction, the discovery pipeline, preparing the data |
| Tuesday | 8.9. | L2: Proximity measures and representations |
| Wednesday | 9.9. | L3: Projection methods, is there any structure at all |
| Tuesday | 15.9. | L4: Partitioning and agglomerative grouping |
| Tuesday | 22.9. | L5: Graph-cut grouping, cluster validity |
| Tuesday | 29.9. | L6: Learned embeddings for grouping, frequency-driven rule search |
| Tuesday | 6.10. | L7: Significance testing for rules |
| Tuesday | 20.10. | L8: Constrained and condensed pattern sets |
| Tuesday | 27.10. | L9: Mining structured and graph data |
| Tuesday | 3.11. | L10: Community structure in networks |
| Tuesday | 10.11. | L11: Stream mining, recommendation |
| Tuesday | 17.11. | L12: Synthesis |
| Tuesday | 24.11. | room booked, no lecture scheduled |
Prerequisites
Expected background: fluency in one programming language at the level of CS-A9110, algorithms and data structures at the level of CS-A9140, the probability and statistics toolkit of MS-A9050, and linear algebra at the level of MS-A9000. MS-C9620 on statistical inference helps but is not assumed. The prerequisite test in the section of the same name shows within minutes which of these you should refresh.
Material
The course follows the textbook R. Lindqvist: Patterns in Large Data Sets. Academic Press, 2021. An electronic copy can be read through the Aalto library (login to aalto-primo). Beyond that there will be a handful of additional readings (linked from the course page). The material belonging to each topic is listed in section Lectures, below the lecture it goes with.
Everything handed out during the term — recordings, slides, exercise sheets, extra notes — lands here in MyCourses.
Workload
Budget roughly 140h in total. Of those, 36h go to contact teaching (12 lectures plus 6 exercise sessions), 18h to the homework and 18h to revising for the exam; the remaining 68h split between exercise work and reading on your own. How that 68h divides is a matter of temperament. Some students read a topic until it is solid and then write the week’s solutions in one sitting, on average 70min per task, which puts them near 44h of reading and 24h of exercises. Others treat the exercise sheet as the way in and consult the material as they go, closer to 34h on each side and about 2h per task. Skipping lectures or sessions moves hours into self-study rather than removing them. Either rhythm works, but reserve a fixed weekly slot for reading regardless: the exercises deliberately do not touch every corner of the syllabus.
Grading
Four things feed into the course performance:
- the prerequisite test, deadline 14.9.2026 (max 2p)
- exercise tasks solved on your own plus taking part in the sessions — 18 tasks across 6 sessions, max 18p
- homework returned by teams of two or three, 4 tasks in all, max 12p
- the exam, Tue 8.12. at 13:00–16:00 (max 28p)
Sum 60p
Your grade follows from the total across those four components. Two thresholds have to be cleared at once: half of the overall points and half of the exam points.
Grade boundaries: grade 1 from 30 points, grade 2 from 36, grade 3 from 42, grade 4 from 48 and grade 5 from 54 upwards
Communication
Anything every participant needs to know goes out as a MyCourses announcement — it shows up on this page and is pushed to everyone enrolled unless you turn that off. Longer conversations, questions and advice belong in the course chat at https://coursechat.example.com/pattern-discovery-2026.
Ask in the chat, at the end of a lecture or in an exercise session — whichever is nearest to hand. Anything that concerns only you is better raised in the weekly staff office hour listed in the syllabus, which keeps the channel readable for everyone else. Asking in the open usually gets you an answer sooner, and it leaves that answer where the next person with the same question will find it.
The opening lecture walks through the rest of the practicalities!