Progress for week 7 (2015)

(Difference between revisions)
 Revision as of 18:37, 15 February 2015 (view source)Elmruud (Talk | contribs) (→Else-Line)← Older edit Current revision as of 09:07, 16 February 2015 (view source)Eivinwi (Talk | contribs) (→Accounting) (5 intermediate revisions not shown) Line 22: Line 22: * Continue improving position-system * Continue improving position-system * Test new grippers, and use them to score points. * Test new grippers, and use them to score points. + + === Accounting === + * Wrote 10 pages + * Helped design new beacon tower + * Tested grippers + * Tested IMU and Color sensors + * No points :'( == Rune == == Rune == Line 30: Line 37: === Accounting === === Accounting === + *Fixed some bugs + *Added/adjusted behaviors/scenario elements + *Started on path-observations == Snorre == == Snorre == Line 37: Line 47: * Write on thesis about the implementation phase, finish the background chapter * Write on thesis about the implementation phase, finish the background chapter === Accounting === === Accounting === + * Made some different transferability functions + * Plotted some results to use in the thesis + * Almost done writing the background chapter == Else-Line == == Else-Line == Line 50: Line 63: ** Both amp-off-phase and koos-sym ** Both amp-off-phase and koos-sym ** Lambda of 2 and 4 with koos-sym ** Lambda of 2 and 4 with koos-sym - * Plotted Baldwinian and Lamarckian results with amp-off-phase. [[plots]] + * Plotted Baldwinian and Lamarckian results with amp-off-phase. ** Baldwinian learning performs worse than no learning, but Lamarckian learning performs better in most cases ** Baldwinian learning performs worse than no learning, but Lamarckian learning performs better in most cases * Wrote some more background material * Wrote some more background material Line 61: Line 74: === Accounting === === Accounting === + * Wrote a better validation and test framework. + * Compared results. Some degree of overfitting. + * Planned further experiments. + * Some text on background and experiments written. == Bendik == == Bendik ==

• Todo 1
• Todo 2

• Done 1
• Done 2

André

Budget

• Have written 20 pages
• Run width tests of beacons, and stability tests
• Create heat maps and statistics of the beacon errors

Eivind

Budget

• Write a minimum of 10 pages
• Continue improving position-system
• Test new grippers, and use them to score points.

Accounting

• Wrote 10 pages
• Helped design new beacon tower
• Tested grippers
• Tested IMU and Color sensors
• No points :'(

Rune

Budget

• Start implementing path observations
• Run some tests
• Write

Accounting

• Fixed some bugs
• Started on path-observations

Snorre

Budget

• Continue working on the transferability function
• Do multi-objective experiments, put the results in tables and graphs
• Write on thesis about the implementation phase, finish the background chapter

Accounting

• Made some different transferability functions
• Plotted some results to use in the thesis
• Almost done writing the background chapter

Else-Line

Budget

• Run more tests
• Plot and interpret results
• Write

Accounting

• Tested out Microsoft Azure Virtual Machines, works pretty well
• Set up second server
• Ran tests
• Both amp-off-phase and koos-sym
• Lambda of 2 and 4 with koos-sym
• Plotted Baldwinian and Lamarckian results with amp-off-phase.
• Baldwinian learning performs worse than no learning, but Lamarckian learning performs better in most cases
• Wrote some more background material

Stian

Budget

• Write!
• Compare results of the simulations finished so far.
• Keep simulating.

Accounting

• Wrote a better validation and test framework.
• Compared results. Some degree of overfitting.
• Planned further experiments.
• Some text on background and experiments written.

• Write..
• See a doctor