Search for: All records

Multi-head-self-attention (MHSA) mechanisms achieve state-of-the-art (SOTA) performance across natural language processing and vision tasks. However, their quadratic dependence on sequence lengths has bottlenecked inference speeds. To circumvent this bottleneck, researchers have proposed various sparse-MHSA models, where a subset of full attention is computed. Despite their promise, current sparse libraries and compilers do not support high-performance implementations fordiversesparse-MHSA patterns due to the underlying sparse formats they operate on. On one end, sparse libraries operate ongeneral sparse formatswhich target extreme amounts of random sparsity (<10% non-zero values) and have high metadata inO(nnzs). On the other end, hand-written kernels operate oncustom sparse formatswhich target specific sparse-MHSA patterns. However, the sparsity patterns in sparse-MHSA are moderately sparse (10-50% non-zero values) and varied, resulting in general sparse formats incurring high metadata overhead and custom sparse formats covering few sparse-MSHA patterns, trading off generality for performance. We bridge this gap, achieving both generality and performance, by proposing a novel sparse format: affine-compressed-sparse-row (ACSR) and supporting code-generation scheme, SPLAT, that generates high-performance implementations for diverse sparse-MHSA patterns on GPUs. Core to our proposed format and code generation algorithm is the observation that common sparse-MHSA patterns have uniquely regular geometric properties. These properties, which can be analyzed just-in-time, expose novel optimizations and tiling strategies that SPLAT exploits to generate high-performance implementations for diverse patterns. To demonstrate SPLAT’s efficacy, we use it to generate code for various sparse-MHSA models, achieving speedups of up-to 2.05x and 4.05x over hand-written kernels written in triton and TVM respectively on A100 GPUs in single-precision. 

We present TIPP3 and TIPP3-fast, new tools for abundance profiling in metagenomic datasets. Like its predecessor, TIPP2, the TIPP3 pipeline uses a maximum likelihood approach to place reads into labeled taxonomies using marker genes, but it achieves superior accuracy to TIPP2 by enabling the use of much larger taxonomies through improved algorithmic techniques. We show that TIPP3 is generally more accurate than leading methods for abundance profiling in two important contexts: when reads come from genomes not already in a public database (i.e., novel genomes) and when reads contain sequencing errors. We also show that TIPP3-fast has slightly lower accuracy than TIPP3, but is also generally more accurate than other leading methods and uses a small fraction of TIPP3’s runtime. Additionally, we highlight the potential benefits of restricting abundance profiling methods to those reads that map to marker genes (i.e., using a filtered marker-gene based analysis), which we show typically improves accuracy. TIPP3 is freely available athttps://github.com/c5shen/TIPP3. 

Contrastive learning is an effective unsupervised method in graph representation learning. The key component of contrastive learning lies in the construction of positive and negative samples. Previous methods usually utilize the proximity of nodes in the graph as the principle. Recently, the data-augmentation-based contrastive learning method has advanced to show great power in the visual domain, and some works have extended this method from images to graphs. However, unlike the data augmentation on images, the data augmentation on graphs is far less intuitive and it is much harder to provide high-quality contrastive samples, which leaves much space for improvement. In this work, by introducing an adversarial graph view for data augmentation, we propose a simple but effective method,Adversarial Graph Contrastive Learning(ArieL), to extract informative contrastive samples within reasonable constraints. We develop a new technique calledinformation regularizationfor stable training and use subgraph sampling for scalability. We generalize our method from node-level contrastive learning to the graph level by treating each graph instance as a super-node.ArieLconsistently outperforms the current graph contrastive learning methods for both node-level and graph-level classification tasks on real-world datasets. We further demonstrate thatArieLis more robust in the face of adversarial attacks.