Edited, memorised or added to reading queue

on 18-Dec-2025 (Thu)

Do you want BuboFlash to help you learning these things? Click here to log in or create user.

Flashcard 7779613805836

Tags
#RNN #ariadne #behaviour #consumer #deep-learning #priority #recurrent-neural-networks #retail #simulation #synthetic-data
Question

RNN details

We use a simple RNN architecture with a single LSTM layer and ten-dimensional cell states. The hidden state at the last time-step is combined with binary non-history features to make the final prediction in a logistic layer. Thus, the final prediction of the RNN is linear in the learned and non-history features. The non-history features describe time, weekday, and behavioral gender and are also provided to the baseline methods. Instead of absolute timestamps, the time differences ∆(x t−1 , x t ) to the previous inputs x t−1 are fed to the RNN at each time- step t. Furthermore, the [...] between the last event x T and the prediction time (the session start) is provided to the final prediction layer

Answer
difference

statusnot learnedmeasured difficulty37% [default]last interval [days]               
repetition number in this series0memorised on               scheduled repetition               
scheduled repetition interval               last repetition or drill

Parent (intermediate) annotation

Open it
lso provided to the baseline methods. Instead of absolute timestamps, the time differences ∆(x t−1 , x t ) to the previous inputs x t−1 are fed to the RNN at each time- step t. Furthermore, the <span>difference between the last event x T and the prediction time (the session start) is provided to the final prediction layer <span>

Original toplevel document (pdf)

cannot see any pdfs






Flashcard 7779889581324

Tags
#ML-engineering #ML_in_Action #learning #machine #software-engineering
Question
Generally, these failures happen because the DS team is either inexperienced with solving a problem of the scale required (a [...] or process-driven failure) or hasn’t fully understood the desired outcome from the business (a communication- driven failure)
Answer
technological

statusnot learnedmeasured difficulty37% [default]last interval [days]               
repetition number in this series0memorised on               scheduled repetition               
scheduled repetition interval               last repetition or drill

Parent (intermediate) annotation

Open it
Generally, these failures happen because the DS team is either inexperienced with solving a problem of the scale required (a technological or process-driven failure) or hasn’t fully understood the desired outcome from the business (a communication- driven failure)

Original toplevel document (pdf)

cannot see any pdfs






Annotation 7779892727052

#ML-engineering #ML_in_Action #learning #machine #software-engineering
Avoiding discussing implementation details at this (very early, initial) stage of the project is not merely critical for the team to focus on the problem. Keeping the esoteric details of algorithms and solution design out of discussions with the larger team keeps the business unit members engaged.
statusnot read reprioritisations
last reprioritisation on suggested re-reading day
started reading on finished reading on


Parent (intermediate) annotation

Open it
ocused on how it will be built keeps the DS members of the group focused on the problem. Ignoring when it will be built by helps the business keep its focus aligned on the needs of the project. <span>Avoiding discussing implementation details at this (very early, initial) stage of the project is not merely critical for the team to focus on the problem. Keeping the esoteric details of algorithms and solution design out of discussions with the larger team keeps the business unit members engaged. <span>

Original toplevel document (pdf)

cannot see any pdfs



Flashcard 7779896134924

Tags
#Linux
Question

How to delete all files before a certain date in Linux

If you have a list of files, but you only want to delete files older the a certain date, for example, a maildir folder with 5 years worth of email, and you want to delete everything older then 2 years, then run the following command.

find . -type f -mtime +XXX -maxdepth 1 -exec rm {} \;

The syntax of this is as follows.

. (dot) – signifies the [...]. You can change this to something like /home/someuser/mail/somedomain/someemail/cur or whatever path you nee

Answer
current folder

statusnot learnedmeasured difficulty37% [default]last interval [days]               
repetition number in this series0memorised on               scheduled repetition               
scheduled repetition interval               last repetition or drill

Parent (intermediate) annotation

Open it
you want to delete everything older then 2 years, then run the following command. find . -type f -mtime +XXX -maxdepth 1 -exec rm {} \; The syntax of this is as follows. . (dot) – signifies the <span>current folder. You can change this to something like /home/someuser/mail/somedomain/someemail/cur or whatever path you nee <span>

Original toplevel document

How to delete all files before a certain date in Linux
How to delete all files before a certain date in Linux Posted on: September 15, 2015 If you have a list of files, but you only want to delete files older the a certain date, for example, a maildir folder with 5 years worth of email, and you want to delete everything older then 2 years, then run the following command. find . -type f -mtime +XXX -maxdepth 1 -exec rm {} \; The syntax of this is as follows. find – the command that finds the files . – the dot signifies the current folder. You can change this to something like /home/someuser/mail/somedomain/someemail/cur or whatever path you need -type f – this means only files. Do not look at or delete folders -mtime +XXX – replace XXX with the number of days you want to go back. for example, if you put -mtime +5, it will delet






Annotation 7779898232076

#feature-engineering #lstm #recurrent-neural-networks #rnn
Several students (and the instructor) included a relative measure of time gap compared to contact's recency. For instance, an average time gap between donations of 300 days, and recency of 150 days, gives a ratio of 0.5. A ratio close to 1 indicates perfect timing for a solicitation. A value above 1 indicates the donor might have churned.
statusnot read reprioritisations
last reprioritisation on suggested re-reading day
started reading on finished reading on


Parent (intermediate) annotation

Open it
Several students (and the instructor) included a relative measure of time gap compared to contact's recency. For instance, an average time gap between donations of 300 days, and recency of 150 days, gives a ratio of 0.5. A ratio close to 1 indicates perfect timing for a solicitation. A value above 1 indicates the donor might have churned. Variants included the introduction of standard deviations in the computations (confidence interval) and mathematical transformations (log-transform, square root.)

Original toplevel document (pdf)

cannot see any pdfs



Flashcard 7779900591372

Tags
#RNN #ariadne #behaviour #consumer #deep-learning #priority #retail #simulation #synthetic-data
Question
Deep Neural Network Ensembles for Time Series Classification [8]. In this paper, authors show how an ensemble of multiple [...] can improve upon the state-of-the-art performance of individual neural networks.
Answer
Convolutional Neural Networks

statusnot learnedmeasured difficulty37% [default]last interval [days]               
repetition number in this series0memorised on               scheduled repetition               
scheduled repetition interval               last repetition or drill

Parent (intermediate) annotation

Open it
Deep Neural Network Ensembles for Time Series Classification [8]. In this paper, authors show how an ensemble of multiple Convolutional Neural Networks can improve upon the state-of-the-art performance of individual neural networks.

Original toplevel document (pdf)

cannot see any pdfs






Flashcard 7779903474956

Question
These so-called [...] or lifelong learning systems, and in particular lifelong deep neural networks (L-DNN), were inspired by brain neurophysiology. These deep learning algorithms separate feature training and rule training and are able to add new rule information on the fly.
Answer
continual

statusnot learnedmeasured difficulty37% [default]last interval [days]               
repetition number in this series0memorised on               scheduled repetition               
scheduled repetition interval               last repetition or drill

Parent (intermediate) annotation

Open it
These so-called continual or lifelong learning systems, and in particular lifelong deep neural networks (L-DNN), were inspired by brain neurophysiology. These deep learning algorithms separate feature training a

Original toplevel document

Deep Learning Has Reinvented Quality Control in Manufacturing—but It Hasn’t Gone Far Enough AI systems that make use of “lifelong learning” techniques are more flexible and faster to train
These so-called continual or lifelong learning systems, and in particular lifelong deep neural networks (L-DNN), were inspired by brain neurophysiology. These deep learning algorithms separate feature training and rule training and are able to add new rule information on the fly. While they still learn features slowly using a large and balanced data set, L-DDNs don't learn rules at this stage. And they don't need images of all known valve defects—the dataset can






Annotation 7779949350156

Commity w git
#git #software-engineering

Zdecydowana większość doświadczonych programistów i zespołów zgodzi się co do jednej odpowiedzi: lepiej mieć więcej mniejszych, logicznie powiązanych commitów.

W świecie inżynierii oprogramowania to podejście nazywa się Atomic Commits (Commity Atomowe).

Oto szczegółowe wyjaśnienie, dlaczego to podejście jest lepsze i jak je stosować w praktyce.

1. Złota zasada: Atomic Commits

Commit powinien być "atomowy", co oznacza, że jest to najmniejsza możliwa jednostka zmiany, która ma sens biznesowy lub techniczny i nie psuje działania aplikacji.

  • Jeden commit = Jedno zadanie logiczne.

  • Jeśli naprawiasz błąd w logowaniu i jednocześnie poprawiasz literówkę w stopce strony – to powinny być dwa osobne commity.

  • Jeśli dodajesz nową funkcjonalność, która wymaga zmian w pliku HTML, CSS i JavaScript – to powinien być jeden commit (ponieważ te zmiany są ze sobą nierozerwalnie związane).

2. Dlaczego małe commity są lepsze?

Podejście "wielka paczka zmian" (tzw. Mega-Commit) jest ryzykowne i utrudnia pracę zespołową. Oto dlaczego małe zmiany wygrywają:

  • Łatwiejsze Code Review: Koledzy z zespołu szybciej sprawdzą 5 commitów po 20 linii kodu, które mają jasne opisy, niż jeden commit z 1000 linii, w którym miesza się refaktoryzacja, naprawa błędu i nowa funkcja.

  • Bezpieczne wycofywanie zmian (git revert): Jeśli wdrożysz 3 zmiany w jednym commicie i jedna z nich okaże się błędna, musisz wycofać całość (również te dwie dobre zmiany). Przy atomowych commitach wycofujesz tylko tę jedną, wadliwą część.

  • Łatwiejsze debugowanie (git bisect): Git posiada narzędzie bisect, które pozwala automatycznie znaleźć commit, który wprowadził błąd. Jeśli commit jest mały, od razu wiesz, gdzie leży przyczyna. Jeśli commit jest ogromny, wiesz tylko, że błąd jest "gdzieś w tych 50 plikach".

  • Przenoszenie zmian (git cherry-pick): Czasami trzeba przenieść konkretną poprawkę na inną gałąź (np. z wersji developerskiej do produkcyjnej jako hotfix). Jeśli poprawka jest "sklejona" z innymi zmianami w dużym commicie, jest to bardzo trudne.

3. Pułapka: Czy "mały" oznacza "zepsuty"?

Nie. To ważne rozróżnienie.

Commitowanie "co 5 minut" tylko po to, by zapisać pracę (gdy kod się nie kompiluje lub testy nie przechodzą), jest złym pomysłem w kontekście publicznej historii (tego, co widzi zespół).

Zasada: Każdy commit na głównej gałęzi (master/main/develop) powinien zostawiać aplikację w stanie działającym.

4. Jak pracować, żeby nie zwariować? (Workflow)

Programiści często boją się małych commitów, bo nie chcą "śmiecić" w historii podczas pracy. Oto rozwiązanie:

  1. Lokalnie: Rób commity tak często, jak chcesz (nawet "work in progress", "save point"). To Twoja brudna robota.

  2. Przed wysłaniem (Push): Użyj Squash (zgnatania commitów).

Przykład:

Robisz zadanie "Dodanie formularza kontaktowego".

...
statusnot read reprioritisations
last reprioritisation on suggested re-reading day
started reading on finished reading on



Annotation 7779951971596

#git #software-engineering

Zdecydowana większość doświadczonych programistów i zespołów zgodzi się co do jednej odpowiedzi: lepiej mieć więcej mniejszych, logicznie powiązanych commitów.

W świecie inżynierii oprogramowania to podejście nazywa się Atomic Commits (Commity Atomowe).
statusnot read reprioritisations
last reprioritisation on suggested re-reading day
started reading on finished reading on

Commity w git
Zdecydowana większość doświadczonych programistów i zespołów zgodzi się co do jednej odpowiedzi: lepiej mieć więcej mniejszych, logicznie powiązanych commitów. W świecie inżynierii oprogramowania to podejście nazywa się Atomic Commits (Commity Atomowe). Oto szczegółowe wyjaśnienie, dlaczego to podejście jest lepsze i jak je stosować w praktyce. 1. Złota zasada: Atomic Commits Commit powinien być "atomowy", co oznacza, że jest to najmni