1. Add to overview [0:30]

2. Revise examples [0:35]

3. Implement better scoring [0:45]

4. Implement faster greedy, look into Kneser-Ney implementation, fix Linux issues [1:45]

5. Profile code + optimize [2:00]

6. Meet with Percy [1:00]

# Monthly Archives: September 2013

# Log 09-16-2013

1. Refactor [3:30]

2. Re-write examples section [1:50]

3. Meet with Percy [0:30]

# Log 09-13-2013

1. Revise EP exposition [1:15]

2. Meet with Huggins [0:45]

3. Lunch with Peter [2:30]

4. Debug experiments [1:50]

# Log 09-12-2013

1. Revise tree DP write-up [0:30]

2. Herding write-up [1:00]

3. Plan out PA write-up [1:00]

4. Write Section 3.3 [2:30]

5. Write factorial HMM exposition [0:20]

6. Re-write -system intro [0:50]

7. Re-work EP with [1:00]

8. Lunch with Ruby [1:00]

# Log 09-11-2013

1. Write up weighted averaging argument [0:45]

2. Determine whether edge-cutting can be viewed as a variational approximation [0:40]

3. Determine whether full construction can be viewed as a variational approximation [0:50]

4. Sketch tree minimization DP [0:10]

5. Write up tree minimization DP [1:30]

6. Revise tree minimization DP [0:30]

7. Throw with Patricia [1:30]

8. Meet with Percy [1:00]

9. Sparse Recovery chalk talk [1:00]

10. Update proofs to take weighted averaging argument into account [1:00]

# Log 09-10-2013

1. E-mail [0:30]

2. Fix herding repo [0:30]

3. Implement smoothing / set up preliminary experiments [3:30]

4. Lunch with Bharath [1:15]

5. Fix weighted averaging argument [2:30]

# Log 09-09-2013

1. Set up NGram code [4:30]

2. Build computer [1:30]