The keynote speech by Vint Cerf addresses the arrival of large-scale language models such as GPT-3 and the risks associated with the use of machine learning. Cerf notes that machine learning is a narrower kind of intelligence that uses multi-layer neural networks, which were previously discounted in the 1960s as inadequate. However, over the last decade or so, they have become extremely powerful in their functionality. Cerf points out that there are risks associated with the use of machine learning, and these risks can be categorized similarly to the risks associated with self-driving cars. Cerf suggests that low-risk applications of machine learning include making suggestions to users, while high-risk applications include medical diagnosis and automated surgery. Cerf notes that one of the challenges associated with machine learning is the inability to predict or anticipate how the system might go off the rails.
Cerf mentions self-driving car levels.
ELIZA was a conversational natural language program from the 1960's. The emacs text editor (used by yours truly) includes ELIZA,
invoked by meta-x doctor. PARRY was another natural language program from the 1970's, built by psychiatrist Kenneth Colby to simulare a paranoid schizophrenic patient.
Vint Cerf decided to have them talk to each other, and published the result as
an
Internet Request for Comments (RFC). Broadly speaking, RFCs document
the underlying technology and development of the Internet.
See RFC 439: PARRY Encounters the DOCTOR.
I hereby solicit suggestions for the video of the day. Please email
me your ideas with explanations. Selected entries will win 5
homework points. If your video is played at the beginning of class,
you must also briefly explain something about the video and
something about yourself - in person.
https://pollev.com/slade
You may also download the app to your phone. Use the "slade" poll id.
The
formation and revision of intuitions, Meyer and Frederick. Cognition, Volume 240, November 2023.
Shane Frederick is a Professor of Marketing at the Yale School of
Management.
The second midterm will similar to the first. You will have 2 hours.
Here is a sample midterm exam. (solutions)
There will also be a question on data structures. For decorators,
there will be a wraps question.
There will be a UNIX question, as in the first midterm. Here is a
sample UNIX transcript
(solutions)
UNIX will cover through principle 4.
The CS 201 midterm
review slides cover UNIX.
PVM.html jupyter notebook - Python Virtual
Machine and hw5.
The current version of Python, 3.11, generates some new byte codes
that were not part of previous versions.
I have added definitions for these op codes to hw5.py.
See dis.html which
includes bytecodes new in 3.11.
Logical Problem of the day
If a bat and a ball together cost $1.10, and the bat costs a dollar
more than the ball, how much does the ball cost ?
Lecture 18: UNIX / Computer Architecture.
Administrivia
Midterm Exam II: Thursday November 9th, 7pm
Computer Architecture
Python ~ (invert) unary operator implements two's complement.
classBinaryOp
Enum for the argument of the BINARY_OP instruction (3.11+).
Arithmetic operations
BinaryOp.ADD (0): x + y BinaryOp.SUBTRACT (10): x - y
BinaryOp.MULTIPLY (5): x * y BinaryOp.TRUE_DIVIDE (11): x / y
BinaryOp.FLOOR_DIVIDE (2): x // y BinaryOp.REMAINDER (6): x % y
BinaryOp.MATRIX_MULTIPLY (4): x @ y BinaryOp.POWER (8): x ** y
Logical and binary operations
BinaryOp.LSHIFT (3): x << y BinaryOp.RSHIFT (9): x >> y BinaryOp.AND
(1): x & y BinaryOp.OR (7): x | y BinaryOp.XOR (12): x ^ y
Inplace operations:
BinaryOp.INPLACE_ADD (13): x += y BinaryOp.INPLACE_SUBTRACT (23): x -=
y BinaryOp.INPLACE_MULTIPLY (18): x *= y BinaryOp.INPLACE_TRUE_DIVIDE
(24): x /= y BinaryOp.INPLACE_FLOOR_DIVIDE (15): x //= y
BinaryOp.INPLACE_REMAINDER (19): x %= y
BinaryOp.INPLACE_MATRIX_MULTIPLY (17): x @= y BinaryOp.INPLACE_POWER
(21): x **= y BinaryOp.INPLACE_LSHIFT (16): x <<= y
BinaryOp.INPLACE_RSHIFT (22): x >>= y BinaryOp.INPLACE_AND (14): x &=
y BinaryOp.INPLACE_OR (20): x |= y BinaryOp.INPLACE_XOR (25): x ^= y
See updated BINARY_OP and SYMBOLS in hw5.py.
Getting to know UNIX
UNIX Introduction Principle 4.
[Home]