...

I was curious for Monero dev's rationale to pick I2P over Tor Whatever I've seen online doesn't strike me as particularly convincing.

Same could be asked of Zcash strong cryptographic ZKP style currencies users often using Tor. As well as a handful of other cryptocurrencies explicitly advertised and designed to use with Tor.

...

Whatever I've seen online doesn't strike me as particularly convincing.

...

Is there published research in regards to deanonymization attacks against both Tor and I2P

Some are here, some are in sites of other messaging systems... https://www.freehaven.net/anonbib All overlay networks currently in production are massively vulnerable to at least two classes of attack by sufficiently interested and capable adversaries... 1) Sybil a) This requires people to actually use PKI to make and use assertions and identities and to punt the results they get from their deep social anal probing of each other in real life as operator peers worldwide... into the consensus, DHT, or whatever mechanism each network uses for node approval and selection. b) Also requires complete ongoing analysis of all known physical and logical metadata and behaviour of the nodes themselves. 2) Global Passive Monitoring The US NSA, Global and Regional Telecom Corporations, and other Entities Worldwide, acting both separately and together, have a complete passive and active view of the internet from at minimum the Global Tier-1 ISP Level, including significant analysis and recording capabilities therein. Yet everyone still stupidly fails to execute at least a few of the seemingly available and reasonable countermeasures... a) Encrypt Everything. Automatic, on by default, strong crypto suites, forward secrecy, tofu, psk, rekeying, whatever works best, etc... both... 1) By and between end to end users, same for server to server... 2) On all physical network links worldwide, every port automagic and independant... fiber, copper, radio, etc... embedded in the network hardware itself via RFC, IEEE, etc b) Deploy fulltime network fill traffic aka chaff, to fill the committed capacity that each node provisioned itself into the [overlay] network with, dynamically yielding room for and upon native traffic. This applies both to, logical nets 2a1, and physical nets 2a2, above. c) Politics, Anarchism, Cryptocurrency Crowdfunding, and whatever else works to uproot and route around persistant known bad actors. 3) Etc Nobody seems to want to do much on the above, to actually shape those into effective global project efforts, to deploy any sufficient mitigation finally therein, therefore the vulnerabilities shall remain. #OpenFabs , #OpenHW , #OpenSW , #OpenDev , #OpenBiz , #CryptoCurrency , #Anarchism , #SybilBusters , #EncryptEverything , #FillEverything ... the list gets longer. Anyone can launch rockets these days. So there is no reason any of the above and more can't be done. Go build and launch some rockets.