APEX Network blockchain development progress

Development status report — #2

Continuing with our new series of more detailed updates on the week to week development work going on in our blockchain division, we are pleased to share with you excerpts from tasks worked on and completed during the weeks 28 & 29.

Week 28 — tasks completed

  • Re-evaluated the code for the APEX Virtual Machine (AVM) module, part of smart contract optimization technology in progress
  • Started working on CPX Ledger app & wallet integration
  • APEX Federal Learning now uses XGBoost to shrink the overall number of leaf nodes. Additionally we optimized our recursive tree algorithm, wich resulted in a significantly smaller memory footprint
  • The communication specifications of the APEX Federal Learning Alliance are being developed
  • Federal Learning technology survey of Jingdong Mathematics
  • Released some code to Github — the code for the APEX Federal Learning module is not public yet for obvious reasons.

Week 29 — tasks completed

  • Set up the dev environment for integration of the Ledger Nano S and the CPX BOLOS app
  • Developed a federal learning protocol that allows the use of raw modeling intermediate data in cases where the data is indivisible. Otherwise encrypted data must be used. This will help users balance encryption and efficiency
  • APEX Federal Learning smart contract design. As part of measures to prevent foul play a certain amount of CPX shall be registered and mortgaged to participate in the APEX Federal Learning Alliance
  • Federal XGBoost regression model optimized, regularization of tree structure during tree construction.

Aiming for a biweekly update schedule, the next update is expected two weeks from now, though this may vary a little depending on what we have going on here. We do hope you appreciate the opportunity to follow these development updates being delivered in more of a granular fashion than before.

Thank you for your time,


APEX Network

Blockchain Powering the Next Generation of Consumer Applications