Source author record

Suman Nath

Suman Nath appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Artificial Intelligence Databases Computation and Language Information Retrieval Operating Systems Software Engineering

Catalog footprint

What is connected

4works

6topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Skim: Speculative Execution for Fast and Efficient Web Agents

Skim is a speculative execution framework for web agents that exploits the predictable structure of purpose-built websites. Today's web-agent expense is not intrinsic to the tasks but a property of how agents are composed: frontier-model inference, browser rendering, and ReAct-style planning are applied to every step of every task regardless of complexity. Skim's key observation is that websites enforce stable URL patterns, answer formats, and task-to-trajectory mappings across queries of the same type, so most queries can bypass these heavyweight components entirely. An offline profiler captures these patterns once per site. At runtime, Skim matches each query to a template, synthesizes the destination URL, and extracts the answer with a small model. A lightweight verifier gates each fast-path output against the query and schema; rare misspeculations cascade to the full agent, warm-started by the fast path's final URL to preserve upstream trajectory progress. Across standard web-agent benchmarks paired with three backboneagents (WebVoyager, AgentOccam, BrowserUse), Skim reduces median per-task cost by 1.9x and latency by 33.4% with no accuracy loss.

preprint2026arXiv

Thinking Ahead: Prospection-Guided Retrieval of Memory with Language Models

Long-horizon personalization requires dialogue assistants to retrieve user-specific facts from extended interaction histories. In practice, many relevant facts often have low semanticsimilarity to the query under dense retrieval. Standard Retrieval-Augmented Generation (RAG) and GraphRAG systems are still largely retrospective: they rely on embedding similarity to the query or on fixed graph traversals, so they often miss facts that matter for the user's needs but lie far from the query in embedding space. Inspired by prospection, the human ability to use imagined futures as cues for recall, we introduce Prospection-Guided Retrieval (PGR), which decouples retrieval from how memories are stored. Given a user query, PGR first expands the goal into a short Tree-of-Thought (ToT) or linear chain of plausible next steps, and uses these steps as retrieval probes rather than relying on the original query alone. The facts retrieved by these probes are then used to personalize the next round of prospection, enabling PGR to uncover additional memories that become relevant only after the simulation is grounded in the user's history. We also introduce MemoryQuest, a challenging multi-session benchmark in which each query is annotated with 3--5 dated reference facts subject to a low query-reference similarity constraint. Across 1,625 queries spanning 185 user profiles from 3 publicly available datasets, PGR-TOT substantially improves retrieval, including nearly 3x recall on MemoryQuest over the strongest baseline. In pairwise LLM-as-judge comparisons against baselines, PGR-generated responses are preferred on 89--98% of queries, with blinded human annotations on held-out subsets showing the same trend. Overall, the results demonstrate that explicit prospection yields large gains in long-horizon retrieval and response quality relative to similarity-only baselines.

preprint2020arXiv

Causality-Guided Adaptive Interventional Debugging

Runtime nondeterminism is a fact of life in modern database applications. Previous research has shown that nondeterminism can cause applications to intermittently crash, become unresponsive, or experience data corruption. We propose Adaptive Interventional Debugging (AID) for debugging such intermittent failures. AID combines existing statistical debugging, causal analysis, fault injection, and group testing techniques in a novel way to (1) pinpoint the root cause of an application's intermittent failure and (2) generate an explanation of how the root cause triggers the failure. AID works by first identifying a set of runtime behaviors (called predicates) that are strongly correlated to the failure. It then utilizes temporal properties of the predicates to (over)-approximate their causal relationships. Finally, it uses fault injection to execute a sequence of interventions on the predicates and discover their true causal relationships. This enables AID to identify the true root cause and its causal relationship to the failure. We theoretically analyze how fast AID can converge to the identification. We evaluate AID with six real-world applications that intermittently fail under specific inputs. In each case, AID was able to identify the root cause and explain how the root cause triggered the failure, much faster than group testing and more precisely than statistical debugging. We also evaluate AID with many synthetically generated applications with known root causes and confirm that the benefits also hold for them.

preprint2013arXiv

Scalable Visibility Color Map Construction in Spatial Databases

Recent advances in 3D modeling provide us with real 3D datasets to answer queries, such as "What is the best position for a new billboard?" and "Which hotel room has the best view?" in the presence of obstacles. These applications require measuring and differentiating the visibility of an object (target) from different viewpoints in a dataspace, e.g., a billboard may be seen from two viewpoints but is readable only from the viewpoint closer to the target. In this paper, we formulate the above problem of quantifying the visibility of (from) a target object from (of) the surrounding area with a visibility color map (VCM). A VCM is essentially defined as a surface color map of the space, where each viewpoint of the space is assigned a color value that denotes the visibility measure of the target from that viewpoint. Measuring the visibility of a target even from a single viewpoint is an expensive operation, as we need to consider factors such as distance, angle, and obstacles between the viewpoint and the target. Hence, a straightforward approach to construct the VCM that requires visibility computation for every viewpoint of the surrounding space of the target, is prohibitively expensive in terms of both I/Os and computation, especially for a real dataset comprising of thousands of obstacles. We propose an efficient approach to compute the VCM based on a key property of the human vision that eliminates the necessity of computing the visibility for a large number of viewpoints of the space. To further reduce the computational overhead, we propose two approximations; namely, minimum bounding rectangle and tangential approaches with guaranteed error bounds. Our extensive experiments demonstrate the effectiveness and efficiency of our solutions to construct the VCM for real 2D and 3D datasets.