| Dates | Topic | Readings | Lecture slides | Reminders
|
| M 8-25 | Philosophy: the empirical way of thinking about language | Pereira, 2000; Abney, 1996; blog post on the science/engineering question - comments are especially interesting |
|
| W 8-27 | Evaluation, experimentation, and hypothesis testing | | pdf |
|
| M 9-1 | Labor day (no class) | |
|
| W 9-3 | Numerical optimization | Dan Klein's tutorial on Lagrangean multipliers (was this helpful? let me know!) | pdf
|
| M 9-8 | Weighted dynamic programming | Goodman, 1999; McAllester, 2002; Eisner, Goldlust, and Smith, 2005; if you're in love, Shieber, Schabes, and Pereira, 1995 (warning: these papers focus to some extent on parsing algorithms, which won't be covered in this lecture much) | pdf | assignment 1 out
|
| W 9-10 | | | pdf | Due: short summary of proposed literature review topic and hyperlinks to 6-8 papers you think are appropriate. Submit via your User page on the wiki, and link to that page from Noah's user page.
|
| M 9-15 | Stochastic models of sequences: Markov models, hidden Markov models, and related algorithms | Manning and Schütze, 1999 (ch. 9); Smith, 2004 | pdf
|
| W 9-17 | | some notes related to forward/backward probabilities | pdf | assignment 1 due
|
| M 9-22 | Log-linear/exponential/maximum entropy models | Log-linear model article draft sent to you by email (sections 1, 2, and 3). | pdf | assignment 2 out
|
| W 9-24 | Training log-linear models, maximum entropy | Three tutorials: Adam Berger's tutorial; Smith, 2004, and Ratnaparkhi, 1997. Chen and Rosenfeld 1999; Rosenfeld, Chen, and Zhu, 2000; Della Pietra, Della Pietra, and Lafferty, 1995
| pdf |
|
| M 9-29 | Conditional random fields (structured log-linear models) | Lafferty, McCallum, and Pereira, 2001 | pdf | assignment 2 due Tuesday 9pm
|
| W 10-1 | Review session with Kevin Gimpel (TA); brief intro to parsing | | pdf |
|
| M 10-6 | Stochastic and weighted context-free grammars, statistical parsing with CFGs | Johnson, 1998 | pdf | This week: literature review progress meetings with instructor. assignment 3 out
|
| W 10-8 | | Charniak, 1997; Charniak, 2000; Collins, 2003; Klein and Manning, 2003 | pdf
|
| M 10-13 | | | pdf
|
| W 10-15 | Other discriminative methods for structured data: perceptron, boosting, maximizing the margin, online methods | Collins, 2002; Altun, Johnson, and Hofmann, 2003; Taskar and Klein's ACL 2005 tutorial | pdf | assignment 3 due (extension: now due F 10-17 at 7am)
|
| M 10-20 | | | pdf
|
| W 10-22 | Dependency parsing (guest lecture - Dipanjan Das and André Martins) | McDonald, Pereira, Ribarov, and Hajic, 2005; Nivre and McDonald, 2008 | pdf | Due: literature review draft (extension: now due F 10-24 at 5pm). assignment 4 out.
|
| M 10-27 | EM, word clustering, and word alignment (guest lecture - Kevin Gimpel) | Brown et al., 1992; Brown et al., 1993 | pdf |
|
| W 10-29 | Bayesian methods in NLP (guest lecture - Shay Cohen) | Blei et al., 2003; Teh et al., 2004 | pdf
|
| M 11-3 | EM for NL models, contrastive estimation | Merialdo, 1994; Pereira and Schabes, 1992; Klein and Manning, 2002; Smith and Eisner, 2005 | pdf
|
| W 11-5 | | | pdf | assignment 4 due
|
| M 11-10 | Bayesian NL models | Blei, Ng, and Jordan, 2003; Liang and Klein's ACL 2007 tutorial | pdf
|
| W 11-12 | Approximate inference for NL models | Goldwater and Griffiths, 2007; Johnson, 2007; Beal, 2003 (ch. 3); MacKay, 1997 | pdf
|
| M 11-17 | Combining labeled and unlabeled data: Yarowsky algorithms, bootstrapping, self-training, co-training | Yarowsky, 1995; Blum and Mitchell, 1998; Nigam and Ghani, 2000; Abney, 2004, Smith and Eisner, 2007, Mann and McCallum, 2007, McClosky et al., 2006; see also Jerry Zhu's semisupervised learning survey | pdf
|
| W 11-19 | | | pdf
|
| M 11-24 | Weighted finite-state machines and transducers | Mohri's list of references on algorithms; Eisner, 2002; Stolcke and Omohundro 1993, Kartunnen, 2001 . Tools: Xerox's FS group; AT&T FSM libraries; RWTH FSA toolkit; OpenFST | pdf | Due: literature review. assignment 5 out
|
| W 11-26 | Thanksgiving break (no class) | |
|
| M 12-1 | Oral presentations: Clark/Gonzalez, Bosaghzadeh/Schneider, Balasubramanyan, Razavian/Zollmann | |
|
| W 12-3 | Oral presentations: Kulkarni/Banerjee, Chaudhuri/Pino, Kang/Yano | | | assignment 5 due Friday 12-5 at 3am
|
| M 12-8 | Final exam (1-4pm, room TBA) | |
|
