Anda di halaman 1dari 3

Computer

 Sec. 171.4 of IP Code: electronic or similar device having information-processing capabilities

Computer Program
 Sec. 171.4 of IP Code: set of instructions expressed in words, codes, schemes or in any other
form, which is capable when incorporated in a medium that the computer can read, or causing
the computer to perform or achieve a particular task or result

Copyright
 How acquired? From moment of creation, but it must be
o original; and
 personal involvement or creativity: “modicum level of creativity”
o an artistic literary work = expression
 vs. functional or technical work

Copyright and Material Object


 Sec. 181 of IP Code as amended: The copyright is distinct from the property in the material object subject
to it. Consequently, the transfer, assignment or licensing of the copyright shall not itself
constitute a transfer of the material object. Nor shall a transfer or assignment of the sole copy
or of one or several copies of the work imply transfer, assignment or licensing of the copyright.

Copyright Infringement
 when liable?
a. unauthorized use of valid copyright
b. unfair use
 who is liable?
a. directly commits an infringement
b. benefits from the infringing activity of another person who commits an infringement if
the person benefiting has been given notice of the infringing activity and has the right and
ability to control the activities of the other person
c. with knowledge of infringing activity, induces, causes or materially contributes to the
infringing conduct of another

Dataset
 in the matter of machine learning, two data sets are used:
o training set
o test set: final random sampling

Economic Rights
1. reproduction
2. dramatization, translation, adaptation, abridgment, arrangement
3. first public distribution of original and each copy of the work
4. rental
5. public display of original or copy of work
6. public performance of work
7. other communication to public of work
Emergent property of AI
 neural network with the ability to refine the “code” in its own subroutines

Hypertext
 text displayed on a computer or other electronic device with references to other text that the
reader can immediately access, usually by a mouse click or keypress sequence

Machine Learning
 involves a learning algorithm, where the algorithm alters itself over time as it is exposed to
data
o self-modication, a.k.a. training, e.g. The algorithm learns how to make better decisions
based on that data, just as a human learn when they start making better decisions based
on experience; e.g. don’t touch the pot of water boiling on the stove, since last time
you burned your hand. (Life lessons… ¯\_(ツ)_/¯)

Moral Rights
 allows the author to take certain actions to preserve the personal link between himself and the
work
1. right of attribution/paternity right (forever)
2. alteration and non-publication right
3. right to preservation of integrity
4. right against false attribution

Parameter
 quantity that helps define a particular process
 a side measure, a numerical characteristic of the algorithm

Procedural content generation


 programmatic generation of game content using a random or pseudo-random process that
results in an unpredictable range of possible game play spaces

Recurrent Neural Network


 a network that cycles information through a loop
 take as their input not just the current input example they see, but also what they have
perceived previously in time
o cf feedforward network (FFN), the only input which it considers is the current example
it has been exposed to
 have memory: there is information in the sequence itself, and RNNs use it to perform tasks
that FFNs can’t
o sequential information is preserved in the recurrent network’s hidden state, which
manages to span many time steps as it cascades forward to affect the processing of
each new examples
o finding correlations between events separated by many moments (called long-term
dependencies)
Training
 involves making guesses about the data, measuring the error in their guesses, and correcting
themselves until they make more accurate guesses
 training set: acts as the benchmark against which deep-learning nets are trained; what they
learn to reconstruct before they’re unleashed on data they haven’t seen before
 output: accurate “model” of data capable of producing good predictions when it is fed new
data that resembles what it trained on
o e.g. If you train a machine-learning model on, say, celebrity faces, it will learn what
humans look like, and with just a little extra learning, you could teach it to transfer
what it knows to photographs of your family and friends, whom it has never seen
before.

Works not protected


a. expired duration of copyright
b. unprotected subject matter
o ideas, procedure, system, method of operation, concept, principle, discovery, and mere
data
o news of the day & mere items of press information
o official text of a legislative, administrative or legal nature
o works of the Government

Anda mungkin juga menyukai