Fairpool is a DEX for personal tokens. Create your own token & receive royalties from trading volume. Increase volume by offering dividends / selling content for your token. Subscribe to our Telegram / Twitter to get notified about updates.
Weekly progress
- ⚒ Worked on a better way to communicate within a team.
- ⚒ Worked on our decision-making system.
Making decisions in Unilang
We have been looking for a way to “provide evidence” for a proposition in Unilang. This is important because it would simplify the communication within the team — if person A wants to prove that their idea is correct, they would need to provide evidence.
Suppose Alice and Bob are working on Fairpool. Alice thinks that it’s better to buy ads from influencers, while Bob thinks that it’s better to buy ads on Telegram channels. This is not an easy decision because it relies on hidden assumptions & partial data. How can they agree on a common strategy?
Options:
- Agree to disagree: divide the budget evenly between Alice and Bob, let them run their campaigns independently.
- Formalize: ask Alice and Bob to formalize their strategies, then compare the strategies by expected gains divided by expected losses.
- Train a neural network, then draw a sample from its action distribution.
We notice that Alice and Bob operate within the same environment. That means their models of environment must be equal. If their models are different, it can be either because:
- Alice and Bob have data which is different and non-conflicting.
- Alice and Bob have data which is different and conflicting.
In the first case, the differences can be resolved simply by merging the data.
In the second case, the differences can be resolved by asking other people to provide more data, or by checking their data against official sources, or by checking their data for internal inconsistencies.
The process of resolving the differences in data is called “reconciliation”.
At any point in time, Alice and Bob can stop the reconciliation and “agree to disagree” (apply their own strategies separately). However, even if stopped, the reconciliation procedure will help both Alice and Bob by at least highlighting the differences (drawing their attention to these data points) and the similarities (reinforcing their belief in these data points).
Reconciliation requires both Alice and Bob to use the same language. That means they need to agree on every definition. We propose a simple formal language for this process, because it allows writing precise definitions.
If Alice and Bob are discussing a marketing strategy, they need to define the behavior of the customer. However, they don’t have complete information about the customer, so they will need to express its behavior in probabilistic terms. However, it’s highly unlikely that Alice and Bob would be familiar with probability theory. So we need a simple way to express their ideas about the possible customer behavior.
We can greatly simplify the burden on Alice and Bob by asking them to only specify the relative probabilities of the state transitions. For example, they could specify the probability of the customer buying a certain item. Then, the system would run multiple simulations to estimate the expected gain / expected loss. By the law of large numbers, the estimate would tend to the actual expected value as the number of simulations would increase.
Such systems are called PTRS — Probabilistic Term Rewriting Systems. We will explore these in more detail & modify Unilang to allow specifying probabilities of the transitions.
Next week’s focus
- ⚒ Develop our decision-making system.
About Fairpool
Fairpool is a DEX for personal tokens. Create your own token & receive royalties from trading volume. Increase volume by offering dividends / selling content for your token. If you want to get notified about updates to our products, please follow our Telegram & Twitter.
Any questions? Reach out to us:
Website: fairpool.io
Telegram: @FairpoolDEX
Twitter: @FairpoolDEX
Anchor Podcasts: @FairpoolDEX
Medium: fairpool-dex.medium.com
$FAIR token: Uniswap
