Live data from Hacker News

Solving the Dating Problem with the SENPAI Protocol [pdf]

sigtbd.csail.mit.edu

31–40 of 113 posts

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#31

Tangential: can somebody explain to me how same-sex speed-dating is sequenced? Two-sex is easy as one sex remains static while the other rotates around, but I can't figure out a solution for single-sex. Update: a solution that satisfies being practical to organise and that lets you talk to all participants Update2: http://math.stackexchange.com/questions/55439/gay-speed-dati...

You divide the people into two groups arbitrarily.

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#32

Tangential: can somebody explain to me how same-sex speed-dating is sequenced? Two-sex is easy as one sex remains static while the other rotates around, but I can't figure out a solution for single-sex. Update: a solution that satisfies being practical to organise and that lets you talk to all participants Update2: http://math.stackexchange.com/questions/55439/gay-speed-dati...

I don't know how they actually do it, but this would work: https://en.wikipedia.org/wiki/Graph_factorization#Complete_g...

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#33

Tangential: can somebody explain to me how same-sex speed-dating is sequenced? Two-sex is easy as one sex remains static while the other rotates around, but I can't figure out a solution for single-sex. Update: a solution that satisfies being practical to organise and that lets you talk to all participants Update2: http://math.stackexchange.com/questions/55439/gay-speed-dati...

A solution: everyone is sequentially numbered. Evens rotate left one round, odds rotate right the next round. That way everyone has to move the same amount (assuming an even number of participants).

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#34

Tangential: can somebody explain to me how same-sex speed-dating is sequenced? Two-sex is easy as one sex remains static while the other rotates around, but I can't figure out a solution for single-sex. Update: a solution that satisfies being practical to organise and that lets you talk to all participants Update2: http://math.stackexchange.com/questions/55439/gay-speed-dati...

A solution: everyone is sequentially numbered. Evens rotate left one round, odds rotate right the next round. That way everyone has to move the same amount (assuming an even number of participants).

If I understand you correctly, not everyone then sees everyone else, which is likely to breed customer dissatisfaction

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#35

Tangential: can somebody explain to me how same-sex speed-dating is sequenced? Two-sex is easy as one sex remains static while the other rotates around, but I can't figure out a solution for single-sex. Update: a solution that satisfies being practical to organise and that lets you talk to all participants Update2: http://math.stackexchange.com/questions/55439/gay-speed-dati...

You divide the people into two groups arbitrarily.

While I suspect that this is true in practice, this is the disadvantage that you don't talk to everyone, even though you can see them.

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#36

> Unfortunately, this protocol is completely dependent on the trustworthiness of Trent, as they receive all of Alice and Bob’s information and are entirely free to manipulate the results of the protocol. If Trent has been compromised, perhaps by a nation-state adversary, Alice and Bob’s love life will be completely in the hands of a malicious attacker Beautifully done.

"nation-state adversary" and "logical AND" both cracked me up. It is surprising to see common themes with my own research on this topic: "thinking too much".

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#37
post #32

Tangential: can somebody explain to me how same-sex speed-dating is sequenced? Two-sex is easy as one sex remains static while the other rotates around, but I can't figure out a solution for single-sex. Update: a solution that satisfies being practical to organise and that lets you talk to all participants Update2: http://math.stackexchange.com/questions/55439/gay-speed-dati...

I don't know how they actually do it, but this would work: https://en.wikipedia.org/wiki/Graph_factorization#Complete_g...

This way you see everyone, but have a logistical nightmare.

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#38
post #5

> In his presentation of the original protocol, Aaronson suggests that zero-knowledge proofs can be used to prevent cheating. However, we find this approach unsatisfying, partly because zero-knowledge proofs would make the protocol much more unwieldy and time-consuming, and partly because we have no clue how to actually implement ZKP in this context I am in love with the honesty in this.

Garbled circuits can be used for the ZKP protocol, and in this case the circuit would just be a single AND gate. https://www.math.ucla.edu/~tdokos/notes_files/garbledCircuit...

Re: Solving the Dating Problem with the SENPAI Protocol [pdf]

#40
This whole paper was great. It starts heavy on the humor but then also delivers just enough actual cryptographic protocol to make the lead-up worth it. This bit at the end was my favorite:

  As is noted in [1], if Alice simply suggests to Bob that
  they carry out a SENPAI exchange, this in and of itself is
  indicative of interest on Alice’s part. Thus, the only nonawkward
  way to actually use the SENPAI protocol is to
  assemble large groups of people, and have every pair of
  people carry out the protocol.
Post reply on HN