More Like this

1. Optimizing JPEG Quantization for Classification Networks

   Li, Zhijing De Sa (January 2020, Resource-Constrained Machine Learning (ReCoML))

   Deep learning for computer vision depends on lossy image compression: it reduces the storage required for training and test data and lowers transfer costs in deployment. Mainstream datasets and imaging pipelines all rely on standard JPEG compression. In JPEG, the degree of quantization of frequency coefficients controls the lossiness: an 8x8 quantization table (Q-table) decides both the quality of the encoded image and the compression ratio. While a long history of work has sought better Q-tables, existing work either seeks to minimize image distortion or to optimize for models of the human visual system. This work asks whether JPEG Q-tables exist that are “better” for specific vision networks and can offer better quality–size trade-offs than ones designed for human perception or minimal distortion. We reconstruct an ImageNet test set with higher resolution to explore the effect of JPEG compression under novel Q-tables. We attempt several approaches to tune a Q-table for a vision task. We find that a simple sorted random sampling method can exceed the performance of the standard JPEG Q-table. We also use hyper-parameter tuning techniques including bounded random search, Bayesian optimization, and composite heuristic optimization methods. The new Q-tables we obtained can improve the compression rate by 10% to 200% when the accuracy is fixed, or improve accuracy up to 2% at the same compression rate. 

   more » « less
2. Optimizing JPEG Quantization for Classification Networks

   Li, Zhijing; De Sa, Christopher; Sampson, Adrian (January 2020, Workshop on Resource-Constrained ML (ReCoML))

   Deep learning for computer vision depends on lossy image compression: it reduces the storage required for training and test data and lowers transfer costs in deployment. Mainstream datasets and imaging pipelines all rely on standard JPEG compression. In JPEG, the degree of quantization of frequency coefficients controls the lossiness: an 88 quantization table (Q-table) decides both the quality of the encoded image and the compression ratio. While a long history of work has sought better Q-tables, existing work either seeks to minimize image distortion or to optimize for models of the human visual system. This work asks whether JPEG Q-tables exist that are “better” for specific vision networks and can offer better quality–size trade-offs than ones designed for human perception or minimal distortion. We reconstruct an ImageNet test set with higher resolution to explore the effect of JPEG compression under novel Q-tables. We attempt several approaches to tune a Q-table for a vision task. We find that a simple sorted random sampling method can exceed the performance of the standard JPEG Q-table. We also use hyper-parameter tuning techniques including bounded random search, Bayesian optimization, and composite heuristic optimization methods. The new Q-tables we obtained can improve the compression rate by 10% to 200% when the accuracy is fixed, or improve accuracy up to 2% at the same compression rate. 

   more » « less
3. Informed Down-Sampled Lexicase Selection: Identifying Productive Training Cases for Efficient Problem Solving

   https://doi.org/10.1162/evco_a_00346  

   Boldi, Ryan; Briesch, Martin; Sobania, Dominik; Lalejini, Alexander; Helmuth, Thomas; Rothlauf, Franz; Ofria, Charles; Spector, Lee (January 2024, Evolutionary Computation)

   Abstract Genetic Programming (GP) often uses large training sets and requires all individuals to be evaluated on all training cases during selection. Random down-sampled lexicase selection evaluates individuals on only a random subset of the training cases, allowing for more individuals to be explored with the same number of program executions. However, sampling randomly can exclude important cases from the down-sample for a number of generations, while cases that measure the same behavior (synonymous cases) may be overused. In this work, we introduce Informed Down-Sampled Lexicase Selection. This method leverages population statistics to build down-samples that contain more distinct and therefore informative training cases. Through an empirical investigation across two different GP systems (PushGP and Grammar-Guided GP), we find that informed down-sampling significantly outperforms random down-sampling on a set of contemporary program synthesis benchmark problems. Through an analysis of the created down-samples, we find that important training cases are included in the down-sample consistently across independent evolutionary runs and systems. We hypothesize that this improvement can be attributed to the ability of Informed Down-Sampled Lexicase Selection to maintain more specialist individuals over the course of evolution, while still benefiting from reduced per-evaluation costs. 

   more » « less
4. Improved Distributed Algorithms for Random Colorings

   https://doi.org/10.4230/LIPICS.OPODIS.2023.13  

   Carlson, Charlie; Frishberg, Daniel; Vigoda, Eric (December 2023, 27th International Conference on Principles of Distributed Systems (OPODIS 2023))

   Bessani, Alysson; Défago, Xavier; Nakamura, Junya; Wada, Koichi; Yamauchi, Yukiko (Ed.)

   Markov Chain Monte Carlo (MCMC) algorithms are a widely-used algorithmic tool for sampling from high-dimensional distributions, a notable example is the equilibirum distribution of graphical models. The Glauber dynamics, also known as the Gibbs sampler, is the simplest example of an MCMC algorithm; the transitions of the chain update the configuration at a randomly chosen coordinate at each step. Several works have studied distributed versions of the Glauber dynamics and we extend these efforts to a more general family of Markov chains. An important combinatorial problem in the study of MCMC algorithms is random colorings. Given a graph G of maximum degree Δ and an integer k ≥ Δ+1, the goal is to generate a random proper vertex k-coloring of G. Jerrum (1995) proved that the Glauber dynamics has O(nlog{n}) mixing time when k > 2Δ. Fischer and Ghaffari (2018), and independently Feng, Hayes, and Yin (2018), presented a parallel and distributed version of the Glauber dynamics which converges in O(log{n}) rounds for k > (2+ε)Δ for any ε > 0. We improve this result to k > (11/6-δ)Δ for a fixed δ > 0. This matches the state of the art for randomly sampling colorings of general graphs in the sequential setting. Whereas previous works focused on distributed variants of the Glauber dynamics, our work presents a parallel and distributed version of the more general flip dynamics presented by Vigoda (2000) (and refined by Chen, Delcourt, Moitra, Perarnau, and Postle (2019)), which recolors local maximal two-colored components in each step. 

   more » « less
5. Multi-Tenant In-memory Key-Value Cache Partitioning Using Efficient Random Sampling-Based LRU Model

   https://doi.org/10.1109/TCC.2023.3300889  

   Wang, Yuchen; Yang, Junyao; Wang, Zhenlin (August 2023, IEEE Transactions on Cloud Computing)

   In-memory key-value caches are widely used as a performance-critical layer in web applications, disk-based storage, and distributed systems. The Least Recently Used (LRU) replacement policy has become the de facto standard in those systems since it exploits workload locality well. However, the LRU implementation can be costly due to the rigid data structure in maintaining object priority, as well as the locks for object order updating. Redis as one of the most effective and prevalent deployed commercial systems adopts an approximated LRU policy, where the least recently used item from a small, randomly sampled set of items is chosen to evict. This random sampling-based policy is lightweight and shows its flexibility. We observe that there can exist a significant miss ratio gap between exact LRU and random sampling-based LRU under different sampling size $$K$$ s. Therefore existing LRU miss ratio curve (MRC) construction techniques cannot be directly applied without loss of accuracy. In this paper, we introduce a new probabilistic stack algorithm named KRR to accurately model random sampling based-LRU, and extend it to handle both fixed and variable objects in key-value caches. We present an efficient stack update algorithm that reduces the expected running time of KRR significantly. To improve the performance of the in-memory multi-tenant key-value cache that utilizes random sampling-based replacement, we propose kRedis, a reference locality- and latency-aware memory partitioning scheme. kRedis guides the memory allocation among the tenants and dynamically customizes $$K$$ to better exploit the locality of each individual tenant. Evaluation results over diverse workloads show that our model generates accurate miss ratio curves for both fixed and variable object size workloads, and enables practical, low-overhead online MRC prediction. Equipped with KRR, kRedis delivers up to a 50.2% average access latency reduction, and up to a 262.8% throughput improvement compared to Redis. Furthermore, by comparing with pRedis, a state-of-the-art design of memory allocation in Redis, kRedis shows up to 24.8% and 61.8% improvements in average access latency and throughput, respectively. 

   more » « less