We lifted 10 functions out of this paper's own repositories and ran 4 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| boschresearch/statestream | canonical | 4 of 10 |
| Function | Status | Where it lives |
|---|---|---|
| S2L | Ran | boschresearch/statestream/statestream/meta/network.py code served (permissive licence) · get_code("cb52df5a86515237") |
| get_item_type | Ran | boschresearch/statestream/statestream/meta/network.py code served (permissive licence) · get_code("62c65ea4bfe94610") |
| get_value | Ran | boschresearch/statestream/statestream/backends/backend_theano.py code served (permissive licence) · get_code("87df0e2f54903130") |
| has_target | Ran | boschresearch/statestream/statestream/meta/losses.py code served (permissive licence) · get_code("ecf1b93e05907804") |
| cm2fd | Not yet run | boschresearch/statestream/statestream/meta/network.py code served (permissive licence) · get_code("43a151508d379182") |
| generate_graph | Not yet run | boschresearch/statestream/statestream/utils/wrapper_networkx.py code served (permissive licence) · get_code("a8c1507b69fb2a97") |
| get_value | Not yet run | boschresearch/statestream/statestream/backends/backend_tensorflow.py code served (permissive licence) · get_code("daebb218663a3f81") |
| import_backend | Not yet run | boschresearch/statestream/statestream/backends/backends.py code served (permissive licence) · get_code("50eda2dd66d22b38") |
| scalar | Not yet run | boschresearch/statestream/statestream/backends/backend_tensorflow.py code served (permissive licence) · get_code("e94f3479871d150e") |
| variable | Not yet run | boschresearch/statestream/statestream/backends/backend_tensorflow.py code served (permissive licence) · get_code("c0eee06fb792d65d") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Deep neural networks, and in particular recurrent networks, are promising candidates to control autonomous agents that interact in real-time with the physical world. However, this requires a seamless integration of temporal features into the network's architecture. For the training of and inference with recurrent neural networks, they are usually rolled out over time, and different rollouts exist. Conventionally during inference, the layers of a network are computed in a sequential manner resulting in sparse temporal integration of information and long response times. In this study, we present a theoretical framework to describe rollouts, the level of model-parallelization they induce, and demonstrate differences in solving specific tasks. We prove that certain rollouts, also for networks with only skip and no recurrent connections, enable earlier and more frequent responses, and show empirically that these early responses have better performance. The streaming rollout maximizes these properties and enables a fully parallel execution of the network reducing runtime on massively parallel devices. Finally, we provide an open-source toolbox to design, train, evaluate, and interact with streaming rollouts.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("1806.04965")
get_code_for_paper("1806.04965")
have("1806.04965")
Connect an agent — have() is free.