osaariki

5 ms·

1.
by osaariki 4mo ago
That 200k is a reasonable amount to start withdrawing from a 5M portfolio (exactly the 4% rule from the Trinity study [1]), but you’ll want to adjust it for inflation every year. My favorite tool for planning these strategies is TPAW Planne
2.
by osaariki 5mo ago
TUIs are great for low friction remote work. I do a lot of data processing work on remote VMs with a mix of interactive debugging/eyeballing and bulk jobs. TUIs are a great fit for the sorts of tools I build to support this work. The o
3.
by osaariki 9mo ago
You're exactly right. The llguidance library [1,2] seems to have emerged as the go-to solution for this by virtue of being >10X faster than its competition. It's work from some past colleagues of mine at Microsoft Research base
4.
by osaariki 11mo ago
I live half way across the world from my folks so I don’t see them often. I’d love something that gives me a greater sense of presence than a video call can give.
5.
by osaariki 2y ago
We don’t know that LLMs generating tokens for scenarios involving simulations of conscious don’t already involve such experience. Certainly such threads of consciousness would currently be much less coherent and fleeting than the human expe
6.
by osaariki 2y ago
For some interesting context: this paper was a precursor to all the work on synthetic data at Microsoft Research that lead to the Phi series of SLMs. [1] It was an important demonstration of what carefully curated and clean data could do fo
7.
by osaariki 2y ago
Edge's Password Monitor feature uses homomorphic encryption to match passwords against a database of leaks without revealing anything about those passwords: https://www.microsoft.com/en-us/research/blog/p
8.
by osaariki 3y ago
I'm not familiar with how TypeChat works, but Guidance [1] is another similar project that can actually integrate into the token sampling to enforce formats. [1]: https://github.com/microsoft/guidance
9.
by osaariki 4y ago
We have an early tool paper [1] for a previous version of the engine, but that's short and with POSIX semantics, so doesn't include a lot of the interesting stuff. The most relevant bit there is the handling of Unicode. >This d
10.
by osaariki 4y ago
We built the new engine behind .NET's RegexOptions.NonBacktracking with derivatives. We will have a paper at PLDI this year on the work that went into that. PCRE semantics was indeed the big thing that required new techniques. Basicall
11.
by osaariki 4y ago
I'd love for someone to do a good quality PyTorch enabled implementation of Sampled AlphaZero/MuZero [1]. RLLib has an AlphaZero, but it doesn't have the parallelized MCTS you really want to have and the "Sampled" p
12.
by osaariki 4y ago
The in-place update work in Koka [1] is super impressive. One of my co-workers, Daan Leijen leads the Koka project and hearing his talks about it have been such a privilege. The work around Koka is really convincing me that functional langu
13.
by osaariki 5y ago
I've done work on low-latency FHE neural network inferencing [1] and we estimated the FLOPS to be approximately 4 times that of an Intel 8087 floating point coprocessor [2]. This was for a LeNet-5 network on the MNIST dataset with mult
14.
by osaariki 5y ago
I found this article confusing too. Being on a tech news site I expected it to eventually get to something concrete, perhaps producing chimeric colors [1] for an impactful moment -- something to justify the "hack your brain" line.
15.
by osaariki 5y ago
The article touches on DFA based engines and notes that they may not be the fastest in all use cases. To expand on that generally the reason for this would be that the automaton you get from the regular expression has too many states, which
16.
by osaariki 6y ago
I do research on FHE. On a high level querying works by evaluating an equivalence function between the encrypted query key and keys in the database such that the result is 1 if the key matches and 0 otherwise. You can then multiply this wit
17.
by osaariki 6y ago
Exactly. And the magic is that now you can securely compute aggregations of data you've stored without having to download all of your data. Say you have some field in all of your encrypted, cloud-stored records and want to find the sum
18.
by osaariki 6y ago
This is not how homomorphic encryptions works. The schemes in use are not deterministic.
19.
by osaariki 6y ago
More than 2^5 has already been demonstrated with GPU implementations, so they definitely do mean 10^5.
20.
by osaariki 6y ago
CPU is currently a rather natural baseline as the various GPU implementations are still somewhat preliminary. I'm not quite an expert on the implementation challenges around FHE (I work on compilers targeting FHE), but my understanding
21.
by osaariki 6y ago
Homomorphic encryption (HE) is NOT what people would understand as "on-chip cryptography", i.e., secure enclaves. HE is a form of encryption that allows arithmetic operations on encrypted numbers without any access to the decrypti
22.
by osaariki 6y ago
Hi! I'm one of the main researchers on this project. If anyone has any questions about the EVA compiler or homomorphic encryption in general I'd be happy to answer.
23.
Microsoft/EVA: Compiler for the SEAL homomorphic encryption library(github.com)
4 points by osaariki 6y ago|1 comments
24.
by osaariki 6y ago
Both BGV and CKKS are so called batched schemes, which allows you to pack multiple values into a single ciphertext (typically thousands) and perform operations on all of the values for the cost of a single homomorphic operation. Batching is
25.
by osaariki 7y ago
You're quite correct that evaluating the back propagation could be done exactly the same as the forward pass. However, if you want to train for more than a handful of steps you'll have to use an operation called bootstrapping to p
26.
by osaariki 7y ago
As other posters already mentioned, there's no control flow in the HE instruction set. This is actually one of the reasons neural networks are a great fit for HE, as they often have fixed data access patterns and no control flow.
27.
by osaariki 7y ago
TF Encrypted has focused more on MPC [1]. At least that's what their arXiv paper [2] talks about. It does seem they are also working to integrate Microsoft SEAL for HE. [1]: https://en.wikipedia.org/wiki/Secure_mul
28.
by osaariki 7y ago
You're right, CryptoNets used a data layout optimized for throughput with a batch size 4096. Since then we've done a lot of work on low latency inference with our CHET compiler [1] and my colleagues with LoLa [2]. It all comes dow
29.
by osaariki 7y ago
This is not how this works. Homomorphic encryption (HE) is best thought of as a software CPU that gives you ways to mutate data (but not read it without the secret key of course). Any computation you run has to be using the instruction set
30.
by osaariki 7y ago
This is wholly untrue. The guarantees HE gives for security of encrypted data is exactly the same as any traditional symmetric or asymmetric crypto system. HE is indeed malleable by design, but doing computation on homomorphically encrypted

More ›