meta user
network profile
| bio | website | |
|---|---|---|
| location | United States | |
| age | 30 | |
| visits | member for | 2 years, 10 months |
| seen | Apr 17 '12 at 2:06 | |
| stats | profile views | 13 |
|
May 30 |
awarded | Yearling |
|
May 30 |
awarded | Teacher |
|
Apr 12 |
answered | Benchmark dataset for decision tree algorithm |
|
Apr 12 |
accepted | Benchmark dataset for decision tree algorithm |
|
Mar 7 |
asked | Benchmark dataset for decision tree algorithm |
|
Feb 2 |
comment |
Reinforcement learning of a policy for multiple actors in large state spaces @tdc, Yes, the agent can move in any direction. I'm actually attempting to discretize the state space by converting "move in any direction and do anything" to "move object X to location Y". 1000 locations may not seem like much, but when you attempt to analyze 10 actors * 50 objects * 1000 locations = 500000 possible actions, all in real-time, it gets to be a difficult task to compute efficiently. |
|
Jan 31 |
comment |
How are classifications merged in an ensemble classifier? My point is that if you take a great classifier and average it's prediction with bad classifiers, you're unlikely to get a better prediction. You're diluting your good prediction. |
|
Jan 25 |
comment |
Reinforcement learning of a policy for multiple actors in large state spaces It's a physical warehouse optimization task. Various robots (i.e. actors) have to decide which of N different objects to move, and to where. By "environmental factors", I mean how close the actor is to the object and each proposed target location, how close the object itself is to the target location, how close the target location is to a wall, etc. They're given a limited amount of time to think and react (so no batch analysis of data), and only get a simple floating-point reward feedback of [0:1]. Not exciting stuff, but it's a massive domain nonetheless. |
|
Jan 24 |
asked | Reinforcement learning of a policy for multiple actors in large state spaces |
|
Jan 22 |
comment |
How are classifications merged in an ensemble classifier? It's not that any classifier is overall "bad" or "good". I'm talking about each classifier's accuracy in specific domains. If one classifier is perfect in a specific domain, then including it in an ensemble may potentially obscure it's usefulness, because all the other classifiers may make bad classifications. |
|
Jan 22 |
asked | How are classifications merged in an ensemble classifier? |
|
Jan 3 |
comment |
Calculating probability of discovery The next state evaluated is always more fit than all the other states in the queue, but it's not necessarily more fit than the most fit states seen in the past. Think of it like someone searching for buried treasure. Say they dig 10 holes, and find \$10 in one of them. Then they dig 10 more holes in the areas adjacent to the one where they found the \$10, but out of those they only find \$5. They then dig 10 more holes in the next adjacent areas, etc. How far should they keep digging before giving up? If they suddenly find \$11, how does that effect their decision to dig? |
|
Jan 3 |
comment |
Calculating probability of discovery @jbowman, Not really. The states are evaluated in the order of fittest-first (i.e. an A-star search), so where a state falls in the sequence is highly dependent on its fitness. |
|
Jan 3 |
asked | Calculating probability of discovery |
|
Sep 27 |
awarded | Scholar |
|
Sep 27 |
accepted | Learning the Structure of a Hierarchical Reinforcement Task |
|
Sep 27 |
awarded | Supporter |
|
Sep 27 |
awarded | Student |
|
Sep 27 |
awarded | Editor |
|
Sep 27 |
revised |
Learning the Structure of a Hierarchical Reinforcement Task edited title |
