CS 440/ECE 448
Fall 2021
Margaret Fleck
Quiz 3 skills list
The quiz will be on Wednesday October 13th, covering material through Computer Vision.
It will be available on moodle 7am to noon central time, and you will have 30 minutes to do it.
Natural Language
Very briefly, who/what/when were the following?
- Julie Lovins
- SHRDLU
- Zork
- Pattern Playback
- Parsey McParseface
Types of systems
- Sample tasks (e.g. translation, question answering, ..)
- Deep vs. shallow
- Examples of shallow systems
- Examples of deep systems
- Danger of seeming too fluent
Processing pipeline
- Speech: waveform, spectrogram, formants, phones, synthesis (formant, concatenative)
- Dividing into words, normalizing words, segmenting words (e.g. suffixes), phonological changes
- Finding morphemes
- Part of speech (POS) tagging
- Parsing
- Constituency vs. dependency trees
- Shallow parsers
- Unlexicalized, class-based, Lexicalized
- Semantics
- Shallow semantics
- Sentiment analysis
- Semantic role labelling
- Co-reference resolution
- Dialog coherency
POS tagging
- General familiarity with common POS tags (e.g. Noun, Determiner)
- Approximate size of typical POS tag sets
- Single word with multiple possible tags
- Baseline tagging algorithm
HMMs
- Baseline algorithm
- Markov assumptions
- Graphical model picture, what variables depend on which other ones in the HMM
- Component probabilities (initial, emission, transition)
- Equations for computing probability of a given tag sequence
- Tag transition possibilities as a finite-state automaton
- At what points does an HMM need smoothing?
- Estimating probabilityof unseen word/tag pairs
- Unseen words, also words seen with novel tags
- Words that appear in development data
- Words in training data that are "hapax," i.e. appear only once
- Open-class vs. closed-class words
- Viterbi (trells) decoding algorithm
- Building the trellis
- Extracting the output tag sequence from the trellis
- Big-O running time
- Extensions: bigram, guessing from word form
Computer vision
Image formation
- Image formation (pinhole camera, real lenses, human eye)
- Digitization (computer, human, including color)
- Edge detection, segmentation
Relating 2D to 3D: why might an object look different in two pictures?
Classification:
- Identifying/naming objects in a picture
- Localizing/registering objects within a picture
- Visual question answering, captioning, semantic role labelling for a picture
Reconstructing 3D geometry
- why is it useful?
- from multiple 2D views of a scene
- from a single picture
Other tasks
- Image generation
- Predicting the future