Downloaded 13 times
Uploaded byJeff Smith
Building Learning Agents
The document discusses the design of reactive learning agents. It describes how agents can utilize sensors and actuators to interact with their environment, as well as knowledge sources and machine learning to guide their functions. Reactive systems are highlighted as they can be responsive, resilient, and elastic by leveraging messaging. The BEAM virtual machine and Elixir language are presented as platforms for building such agents, and various testing and validation techniques are outlined.
![[DSC DACH 24] Increasing user adoption with GenAI offerings - Martin Flechl](https://cdn.slidesharecdn.com/ss_thumbnails/martinflechl-240916182349-4779159b-thumbnail.jpg?width=640&height=640&fit=bounds)
Building Learning Agents
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20. Reactive Strategies Replication ContainmentSupervision
- 21.
- 22. Reactive Machine Learning UncertainDataInfinite Data Laziness Pure Functions Immutable Facts Possible Worlds
- 23.
- 24.
- 25.
- 29.
- 30.
- 31. Source of AIComponents
- 32. Source of AIComponents Yours
- 33. Source of AIComponents Other People’s Yours
- 34.
- 35. Sensors •Email/Messaging •Phone cameras &sensors •Wearables •Speaker mIcs
- 36. Sensors •Email/Messaging •Phone cameras &sensors •Wearables •Speaker mIcs •Bespoke Data Collection Systems
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43. Reactive Data Collection •Replicationvia Distributed databases •Event-sourcing/immutable fact databases •Uncertainty-aware data models •AP Databases https://skillsmatter.com/skillscasts/7038-lightning-talks-2
- 44.
- 45.
- 46. Actuators •Email/Messaging •Notifications •Commerce APIs •Speaker devices •Self-drivingCars •External API Calls •Internal API Calls
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53. Knowledge •External Datasets •Libraries •Knowledge asa Service •Your Training corpus •Libraries •Institutional knowledge
- 54.
- 55. Erlang & theBEAM •Functional Language •Concurrency-oriented •Actor Model •Over 30 years old!
- 56.
- 57. Reactive Knowledge Systems Adaptedfrom https://github.com/mitchellhenke/circuit_breaker
- 58.
- 59.
- 60.
- 61.
- 62.
- 63.
- 64. Learning •Machine Learning as aService APIs •Open Source Libraries
- 65. Learning •Machine Learning as aService APIs •Open Source Libraries •Your novel Algorithms •Organizational Learning
- 66.
- 67.
- 68.
- 69.
- 70.
- 71.
- 72.
- 73.
- 74.
- 75.
- 76.
- 77. Agent Function •Templates •DecisionComponent
- 78.
- 79.
- 80.
- 81.
- 82.
- 83.
- 84.
- 85. Dialyzer in Action ml_system.ex:3:Function predict/1 has no local return ml_system.ex:6: The call 'Elixir.MLSystem':call_model_b(feature@1::number()) will never return since it differs in the 1st argument from the success typing arguments: (binary()) ml_system.ex:22: Invalid type specification for function 'Elixir.MLSystem':call_model_b/1. The success typing is (binary()) -> binary() ml_system.ex:23: Function call_model_b/1 has no local return
- 86. Dialyzer in Action ml_system.ex:24:The call 'Elixir.String':upcase(feature@1::number()) will never return since the success typing is (binary()) -> bitstring() and the contract is (t()) -> t() ml_system.ex:38: Function ensemble/1 will never be called
- 87.
- 88.
- 89.
- 90.
- 91.
- 92.
- 93.
- 94.
- 95. Concuerror Output Interleaving #1 ---------------------------------------------------------------- ------------- Errorsfound: * Blocked at a 'receive' (when all other processes have exited): P.1 in gen_server.erl line 382 ---------------------------------------------------------------- -------------
- 96.
- 97.
- 98.
- 99. Testing Agents onthe BEAM Tool Tests Dialyzer Types QuickCheck Properties Concuerror Interleavings
- 100.
- 101. Agent Frameworks •API.AI (Google) •Lex(Amazon) •Microsoft Bot Framework •Botkit Studio •Recime
- 102.
- 103.
- 104.
- 105.
- 106. Reactive Machine Learning UncertainDataInfinite Data Laziness Pure Functions Immutable Facts Possible Worlds
- 107. Source of AIComponents Other People’s Yours
- 108. Elixir & theBEAM
- 109.
- 110.
- 111.
- 112.
- 113.