Stream-based retrieval augmented generation (RAG) systems integrate stream processing engines (SPEs) with real-time document retrieval to support dynamic indexing and search over unstructured datasets. However, efficiently executing queries in these systems is challenging, as seamless coordination between SPEs and vector databases is essential for maintaining low latency and high throughput. The lack of mutual awareness between these components results in two major performance bottlenecks. First, SPEs are unaware of ongoing indexing operations in vector databases, leading to metadata lock contention when indexing and search operations overlap, which increases query latency. Second, vector databases lack visibility into query traffic from SPEs and rely solely on internal metrics for scaling. As a result, they respond reactively to traffic spikes, often leading to instance overload and delayed query processing. To address these issues, we propose StreamRAG, a lock-aware and traffic-aware query coordination mechanism that facilitates real-time exchange of metadata lock statuses and query traffic metrics between SPEs and vector databases. By optimizing query routing and enabling proactive instance scaling, StreamRAG enhances the performance and scalability of real-time RAG systems. Experimental results demonstrate that Streamrag reduces tail latency by up to$4 \times$at the 99 th percentile and significantly improves overall system performance under varying traffic conditions.

Embedding models capture both semantic and syntactic structures of queries, often mapping different queries to similar regions in vector space. This results in nonuniform cluster access patterns in modern disk-based vector databases. While existing approaches optimize individual queries, they overlook the impact of cluster access patterns, failing to account for the locality effects of queries that access similar clusters. This oversight increases cache miss penalty. To minimize the cache miss penalty, we propose CALL, a context-aware query grouping mechanism that organizes queries based on shared cluster access patterns. Additionally, CALL incorporates a group-aware prefetching method to minimize cache misses during transitions between query groups and latency-aware cluster loading. Experimental results show that CALL reduces the 99th percentile tail latency by up to 33 % while consistently maintaining a higher cache hit ratio, substantially reducing search latency.

The liquidity of the exchange is the most important factor when crytocurrency traders choose an exchange. However, the amount of liquidity provided by the liquidity providers in decentralized exchanges is insufficient when compared to centralized exchanges. This is because the liquidity providers in decentralized exchanges suffer from the risk of divergence loss inherent to the automated market making system. To this end, we introduce a new concept called margin liquidity and leverage this concept to propose a highly profitable margin liquidity-providing position. Then, we extend this margin liquidity-providing position to a virtual margin liquidity-providing position to alleviate the risk of divergence loss for the liquidity providers and encourage them to provide more liquidity to the pool. We show that our proposed margin liquidity is 8K times more capital efficient than the concentrated liquidity proposed in Uniswap V3.

Recent works on neural network pruning advocate that reducing the depth of the network is more effective in reducing run-time memory usage and accelerating inference latency than reducing the width of the network through channel pruning. In this regard, some recent works propose depth compression algorithms that merge convolution layers. However, the existing algorithms have a constricted search space and rely on human-engineered heuristics. In this paper, we propose a novel depth compression algorithm which targets general convolution operations. We propose a subset selection problem that replaces inefficient activation layers with identity functions and optimally merges consecutive convolution operations into shallow equivalent convolution operations for efficient end-to-end inference latency. Since the proposed subset selection problem is NP-hard, we formulate a surrogate optimization problem that can be solved exactly via two-stage dynamic programming within a few seconds. We evaluate our methods and baselines by TensorRT for a fair inference latency comparison. Our method outperforms the baseline method with higher accuracy and faster inference speed in MobileNetV2 on the ImageNet dataset. Specifically, we achieve $1.41\times$ speed-up with $0.11$%p accuracy gain in MobileNetV2-1.0 on the ImageNet.

Reducing the high computational cost of large convolutional neural networks is crucial when deploying the networks to resource-constrained environments. We first show the greedy approach of recent channel pruning methods ignores the inherent quadratic coupling between channels in the neighboring layers and cannot safely remove inactive weights during the pruning procedure. Furthermore, due to these inactive weights, the greedy methods cannot guarantee to satisfy the given resource constraints and deviate with the true objective. In this regard, we propose a novel channel selection method that optimally selects channels via discrete QCQP, which provably prevents any inactive weights and guarantees to meet the resource constraints tightly in terms of FLOPs, memory usage, and network size. We also propose a quadratic model that accurately estimates the actual inference time of the pruned network, which allows us to adopt inference time as a resource constraint option. Furthermore, we generalize our method to extend the selection granularity beyond channels and handle non-sequential connections. Our experiments on CIFAR-10 and ImageNet show our proposed pruning method outperforms other fixed-importance channel pruning methods on various network architectures.

We develop hierarchically quantized efficient embedding representations for similarity-based search and show that this representation provides not only the state of the art performance on the search accuracy but also provides several orders of speed up during inference. The idea is to hierarchically quantize the representation so that the quantization granularity is greatly increased while maintaining the accuracy and keeping the computational complexity low. We also show that the problem of finding the optimal sparse compound hash code respecting the hierarchical structure can be optimized in polynomial time via minimum cost flow in an equivalent flow network. This allows us to train the method end-to-end in a mini-batch stochastic gradient descent setting. Our experiments on Cifar100 and ImageNet datasets show the state of the art search accuracy while providing several orders of magnitude search speedup respectively over exhaustive linear search over the dataset.

We address the problem of unsupervised disentanglement of discrete and continuous explanatory factors of data. We first show a simple procedure for minimizing the total correlation of the continuous latent variables without having to use a discriminator network or perform importance sampling, via cascading the information flow in the beta-VAE framework. Furthermore, we propose a method which avoids offloading the entire burden of jointly modeling the continuous and discrete factors to the variational encoder by employing a separate discrete inference procedure. This leads to an interesting alternating minimization problem which switches between finding the most likely discrete configuration given the continuous factors and updating the variational encoder based on the computed discrete factors. Experiments show that the proposed method clearly disentangles discrete factors and significantly outperforms current disentanglement methods based on the disentanglement score and inference network classification score. The source code is available at https://github.com/snumllab/DisentanglementICML19.

Reinforcement learning algorithms struggle when the reward signal is very sparse. In these cases, naive random exploration methods essentially rely on a random walk to stumble onto a rewarding state. Recent works utilize intrinsic motivation to guide the exploration via generative models, predictive forward models, or discriminative modeling of novelty. We propose EMI, which is an exploration method that constructs embedding representation of states and actions that does not rely on generative decoding of the full observation but extracts predictive signals that can be used to guide exploration based on forward prediction in the representation space. Our experiments show competitive results on challenging locomotion tasks with continuous control and on image-based exploration tasks with discrete actions on Atari. The source code is available at https://github.com/snu-mllab/EMI.

Embedding representation learning via neural networks is at the core foundation of modern similarity based search. While much effort has been put in developing algorithms for learning binary hamming code representations for search efficiency, this still requires a linear scan of the entire dataset per each query and trades off the search accuracy through binarization. To this end, we consider the problem of directly learning a quantizable embedding representation and the sparse binary hash code end-to-end which can be used to construct an efficient hash table not only providing significant search reduction in the number of data but also achieving the state of the art search accuracy outperforming previous state of the art deep metric learning methods. We also show that finding the optimal sparse binary hash code in a mini-batch can be computed exactly in polynomial time by solving a minimum cost flow problem. Our results on Cifar-100 and on ImageNet datasets show the state of the art search accuracy in precision@k and NMI metrics while providing up to 98X and 478X search speedup respectively over exhaustive linear search. The source code is available at https://github.com/maestrojeong/Deep-Hash-Table-ICML18.