Search papers, labs, and topics across Lattice.
This paper introduces Flash-SemiCRF, a novel approach to Semi-Markov Conditional Random Fields (semi-CRFs) that addresses the memory bottleneck of existing implementations by evaluating edge potentials on-the-fly using a compact prefix-sum array. By replacing the large edge potential tensor with this efficient lookup method, the authors achieve a significant reduction in memory usage while maintaining exact inference capabilities for long sequences and large label sets. The proposed streaming forward-backward pass further optimizes memory usage, enabling the application of semi-CRFs to previously intractable problems in speech and genomics.
Flash-SemiCRF slashes memory requirements for segment-level inference, making it feasible for genomic sequences over 100,000 positions.
Semi-Markov Conditional Random Fields (semi-CRFs) assign labels to segments of a sequence rather than to individual positions, enabling exact inference over segment-level features and principled uncertainty estimates at their boundaries. However, existing implementations must materialize a large edge potential tensor whose size grows with sequence length, maximum segment length, and label count, becoming prohibitive for speech-scale state spaces and intractable at genomic scales where sequences can exceed 100,000 positions. This memory bottleneck has limited the adoption of exact segment-level inference for long sequences and large label sets. We identify that the core inefficiency is materializing edge potentials that can instead be evaluated on-the-fly from a compact prefix-sum array, and make several improvements. First, replacing the stored edge tensor with prefix-sum lookup reduces the memory footprint by a factor proportional to the product of segment length and label count. Second, a streaming forward-backward pass with checkpoint-boundary normalization keeps working memory sublinear in sequence length while preserving exact gradients. Third, zero-centered cumulative scores control numerical drift and induce an adaptive duration prior under label imbalance. We integrate these ideas into Flash-SemiCRF, a fused Triton kernel that enables exact semi-CRF inference on previously intractable problem sizes. Available at https://github.com/biobenkj/flash-semicrf.
