A Sparser Johnson-Lindenstrauss Transform

نویسندگان

  • Daniel M. Kane
  • Jelani Nelson
چکیده

We give a Johnson-Lindenstrauss transform with column sparsity s = Θ(ε−1 log(1/δ)) into optimal dimension k = O(ε−2 log(1/δ)) to achieve distortion 1±ε with success probability 1−δ. This is the first distribution to provide an asymptotic improvement over the Θ(k) sparsity bound for all values of ε, δ. Previous work of [Dasgupta-Kumar-Sarlós, STOC 2010] gave a distribution with s = Õ(ε−1 log(1/δ))1, with tighter analyses later in [Kane-Nelson, CoRR abs/1006.3585] and [Braverman-Ostrovsky-Rabani, CoRR abs/1011.2590] showing that their construction achieves s = Õ(ε−1 log(1/δ)). As in the previous work, our scheme only requires limited independence hash functions. In fact, potentially one of our hash functions could be made deterministic given an explicit construction of a sufficiently good error-correcting code. Our linear dependence on log(1/δ) in the sparsity allows us to plug our construction into algorithms of [Clarkson-Woodruff, STOC 2009] to achieve the fastest known streaming algorithms for numerical linear algebra problems such as approximate linear regression and best rank-k approximation. Their reductions to the Johnson-Lindenstrauss lemma require exponentially small δ, and thus a superlinear dependence on log(1/δ) in s leads to significantly slower algorithms.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

The Fast Johnson-lindenstrauss Transform

While we omit the proof, we remark that it is constructive. Specifically, A is a linear map consisting of random projections onto subspaces of Rd. These projections can be computed by n matrix multiplications, which take time O(nkd). This is fast enough to make the Johnson-Lindenstrauss transform (JLT) a practical and widespread algorithm for dimensionality reduction, which in turn motivates th...

متن کامل

A Derandomized Sparse Johnson-Lindenstrauss Transform

Recent work of [Dasgupta-Kumar-Sarlós, STOC 2010] gave a sparse Johnson-Lindenstrauss transform and left as a main open question whether their construction could be efficiently derandomized. We answer their question affirmatively by giving an alternative proof of their result requiring only bounded independence hash functions. Furthermore, the sparsity bound obtained in our proof is improved. T...

متن کامل

Some Useful Background for Talk on the Fast Johnson-Lindenstrauss Transform

This writeup includes very basic background material for the talk on the Fast Johnson Lindenstrauss Transform on May 28th, 2007 by the author at the Summer School on Algorithmic Data Analysis in Helsinki, Finland. The talk is about the results in [2], but will be based on a new more recent proof (together with Edo Liberty), using more modern tools, such as probability in Banach spaces and error...

متن کامل

Optimal Fast Johnson-Lindenstrauss Embeddings for Large Data Sets

We introduce a new fast construction of a Johnson-Lindenstrauss matrix based on the composition of the following two embeddings: A fast construction by the second author joint with Ward [1] maps points into a space of lower, but not optimal dimension. Then a subsequent transformation by a dense matrix with independent entries reaches an optimal embedding dimension. As we show in this note, the ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:
  • CoRR

دوره abs/1012.1577  شماره 

صفحات  -

تاریخ انتشار 2010