Searcharxiv⌕ Search

arXiv subjects

Vinayak Gupta

Publications and source records attributed to Vinayak Gupta.

34 records · Page 2Linked to original sources
arXiv2024

SPML: A DSL for Defending Language Models Against Prompt Attacks

Large language models (LLMs) have profoundly transformed natural language applications, with a growing reliance on instruction-based definitions for designing chatbots. However, post-deployment the chatbot definitions are fixed and are vulnerable to attacks by malicious users, emphasizing the need to prevent unethical applications and financial losses. Existing studies explore user prompts' impact on LLM-based chatbots, yet practical methods to contain attacks on application-specific chatbots remain unexplored. This paper presents System Prompt Meta Language (SPML), a domain-specific language for refining prompts and monitoring the inputs to the LLM-based chatbots. SPML actively checks attack prompts, ensuring user inputs align with chatbot definitions to prevent malicious execution on the LLM backbone, optimizing costs. It also streamlines chatbot definition crafting with programming language capabilities, overcoming natural language design challenges. Additionally, we introduce a groundbreaking benchmark with 1.8k system prompts and 20k user inputs, offering the inaugural language and benchmark for chatbot definition evaluation. Experiments across datasets demonstrate SPML's proficiency in understanding attacker prompts, surpassing models like GPT-4, GPT-3.5, and LLAMA. Our data and codes are publicly available at: https://prompt-compiler.github.io/SPML/.

cs.LG↗
arXiv2024

GSN: Generalisable Segmentation in Neural Radiance Field

Traditional Radiance Field (RF) representations capture details of a specific scene and must be trained afresh on each scene. Semantic feature fields have been added to RFs to facilitate several segmentation tasks. Generalised RF representations learn the principles of view interpolation. A generalised RF can render new views of an unknown and untrained scene, given a few views. We present a way to distil feature fields into the generalised GNT representation. Our GSN representation generates new views of unseen scenes on the fly along with consistent, per-pixel semantic features. This enables multi-view segmentation of arbitrary new scenes. We show different semantic features being distilled into generalised RFs. Our multi-view segmentation results are on par with methods that use traditional RFs. GSN closes the gap between standard and generalisable RF methods significantly. Project Page: https://vinayak-vg.github.io/GSN/

cs.CV↗
arXiv2023

Retrieving Continuous Time Event Sequences using Neural Temporal Point Processes with Learnable Hashing

Temporal sequences have become pervasive in various real-world applications. Consequently, the volume of data generated in the form of continuous time-event sequence(s) or CTES(s) has increased exponentially in the past few years. Thus, a significant fraction of the ongoing research on CTES datasets involves designing models to address downstream tasks such as next-event prediction, long-term forecasting, sequence classification etc. The recent developments in predictive modeling using marked temporal point processes (MTPP) have enabled an accurate characterization of several real-world applications involving the CTESs. However, due to the complex nature of these CTES datasets, the task of large-scale retrieval of temporal sequences has been overlooked by the past literature. In detail, by CTES retrieval we mean that for an input query sequence, a retrieval system must return a ranked list of relevant sequences from a large corpus. To tackle this, we propose NeuroSeqRet, a first-of-its-kind framework designed specifically for end-to-end CTES retrieval. Specifically, NeuroSeqRet introduces multiple enhancements over standard retrieval frameworks and first applies a trainable unwarping function on the query sequence which makes it comparable with corpus sequences, especially when a relevant query-corpus pair has individually different attributes. Next, it feeds the unwarped query sequence and the corpus sequence into MTPP-guided neural relevance models. We develop four variants of the relevance model for different kinds of applications based on the trade-off between accuracy and efficiency. We also propose an optimization framework to learn binary sequence embeddings from the relevance scores, suitable for the locality-sensitive hashing. Our experiments show the significant accuracy boost of NeuroSeqRet as well as the efficacy of our hashing mechanism.

cs.LG↗
arXiv2023

Tapestry of Time and Actions: Modeling Human Activity Sequences using Temporal Point Process Flows

Human beings always engage in a vast range of activities and tasks that demonstrate their ability to adapt to different scenarios. Any human activity can be represented as a temporal sequence of actions performed to achieve a certain goal. Unlike the time series datasets extracted from electronics or machines, these action sequences are highly disparate in their nature -- the time to finish a sequence of actions can vary between different persons. Therefore, understanding the dynamics of these sequences is essential for many downstream tasks such as activity length prediction, goal prediction, next action recommendation, etc. Existing neural network-based approaches that learn a continuous-time activity sequence (or CTAS) are limited to the presence of only visual data or are designed specifically for a particular task, i.e., limited to next action or goal prediction. In this paper, we present ProActive, a neural marked temporal point process (MTPP) framework for modeling the continuous-time distribution of actions in an activity sequence while simultaneously addressing three high-impact problems -- next action prediction, sequence-goal prediction, and end-to-end sequence generation. Specifically, we utilize a self-attention module with temporal normalizing flows to model the influence and the inter-arrival times between actions in a sequence. In addition, we propose a novel addition over the ProActive model that can handle variations in the order of actions, i.e., different methods of achieving a given goal. We demonstrate that this variant can learn the order in which the person or actor prefers to do their actions. Extensive experiments on sequences derived from three activity recognition datasets show the significant accuracy boost of ProActive over the state-of-the-art in terms of action and goal prediction, and the first-ever application of end-to-end action sequence generation.

cs.CV↗
arXiv2022

Modeling Time-Series and Spatial Data for Recommendations and Other Applications

With the research directions described in this thesis, we seek to address the critical challenges in designing recommender systems that can understand the dynamics of continuous-time event sequences. We follow a ground-up approach, i.e., first, we address the problems that may arise due to the poor quality of CTES data being fed into a recommender system. Later, we handle the task of designing accurate recommender systems. To improve the quality of the CTES data, we address a fundamental problem of overcoming missing events in temporal sequences. Moreover, to provide accurate sequence modeling frameworks, we design solutions for points-of-interest recommendation, i.e., models that can handle spatial mobility data of users to various POI check-ins and recommend candidate locations for the next check-in. Lastly, we highlight that the capabilities of the proposed models can have applications beyond recommender systems, and we extend their abilities to design solutions for large-scale CTES retrieval and human activity prediction. A significant part of this thesis uses the idea of modeling the underlying distribution of CTES via neural marked temporal point processes (MTPP). Traditional MTPP models are stochastic processes that utilize a fixed formulation to capture the generative mechanism of a sequence of discrete events localized in continuous time. In contrast, neural MTPP combine the underlying ideas from the point process literature with modern deep learning architectures. The ability of deep-learning models as accurate function approximators has led to a significant gain in the predictive prowess of neural MTPP models. In this thesis, we utilize and present several neural network-based enhancements for the current MTPP frameworks for the aforementioned real-world applications.

cs.IR↗
arXiv2022

Modeling Spatial Trajectories using Coarse-Grained Smartphone Logs

Current approaches for points-of-interest (POI) recommendation learn the preferences of a user via the standard spatial features such as the POI coordinates, the social network, etc. These models ignore a crucial aspect of spatial mobility -- every user carries their smartphones wherever they go. In addition, with growing privacy concerns, users refrain from sharing their exact geographical coordinates and their social media activity. In this paper, we present REVAMP, a sequential POI recommendation approach that utilizes the user activity on smartphone applications (or apps) to identify their mobility preferences. This work aligns with the recent psychological studies of online urban users, which show that their spatial mobility behavior is largely influenced by the activity of their smartphone apps. In addition, our proposal of coarse-grained smartphone data refers to data logs collected in a privacy-conscious manner, i.e., consisting only of (a) category of the smartphone app and (b) category of check-in location. Thus, REVAMP is not privy to precise geo-coordinates, social networks, or the specific application being accessed. Buoyed by the efficacy of self-attention models, we learn the POI preferences of a user using two forms of positional encodings -- absolute and relative -- with each extracted from the inter-check-in dynamics in the check-in sequence of a user. Extensive experiments across two large-scale datasets from China show the predictive prowess of REVAMP and its ability to predict app- and POI categories.

cs.LG↗
arXiv2022

Modeling Continuous Time Sequences with Intermittent Observations using Marked Temporal Point Processes

A large fraction of data generated via human activities such as online purchases, health records, spatial mobility etc. can be represented as a sequence of events over a continuous-time. Learning deep learning models over these continuous-time event sequences is a non-trivial task as it involves modeling the ever-increasing event timestamps, inter-event time gaps, event types, and the influences between different events within and across different sequences. In recent years neural enhancements to marked temporal point processes (MTPP) have emerged as a powerful framework to model the underlying generative mechanism of asynchronous events localized in continuous time. However, most existing models and inference methods in the MTPP framework consider only the complete observation scenario i.e. the event sequence being modeled is completely observed with no missing events -- an ideal setting that is rarely applicable in real-world applications. A recent line of work which considers missing events while training MTPP utilizes supervised learning techniques that require additional knowledge of missing or observed label for each event in a sequence, which further restricts its practicability as in several scenarios the details of missing events is not known apriori. In this work, we provide a novel unsupervised model and inference method for learning MTPP in presence of event sequences with missing events. Specifically, we first model the generative processes of observed events and missing events using two MTPP, where the missing events are represented as latent random variables. Then, we devise an unsupervised training method that jointly learns both the MTPP by means of variational inference. Such a formulation can effectively impute the missing data among the observed events and can identify the optimal position of missing events in a sequence.

cs.LG↗
arXiv2022

ProActive: Self-Attentive Temporal Point Process Flows for Activity Sequences

Any human activity can be represented as a temporal sequence of actions performed to achieve a certain goal. Unlike machine-made time series, these action sequences are highly disparate as the time taken to finish a similar action might vary between different persons. Therefore, understanding the dynamics of these sequences is essential for many downstream tasks such as activity length prediction, goal prediction, etc. Existing neural approaches that model an activity sequence are either limited to visual data or are task specific, i.e., limited to next action or goal prediction. In this paper, we present ProActive, a neural marked temporal point process (MTPP) framework for modeling the continuous-time distribution of actions in an activity sequence while simultaneously addressing three high-impact problems -- next action prediction, sequence-goal prediction, and end-to-end sequence generation. Specifically, we utilize a self-attention module with temporal normalizing flows to model the influence and the inter-arrival times between actions in a sequence. Moreover, for time-sensitive prediction, we perform an early detection of sequence goal via a constrained margin-based optimization procedure. This in-turn allows ProActive to predict the sequence goal using a limited number of actions. Extensive experiments on sequences derived from three activity recognition datasets show the significant accuracy boost of ProActive over the state-of-the-art in terms of action and goal prediction, and the first-ever application of end-to-end action sequence generation.

cs.CV↗
arXiv2022

Inverse formula for distance matrices of gear graphs

Distance matrices of some star like graphs are investigated in \cite{JAK}. These graphs are trees which are stars, wheel graphs, helm graphs and gear graphs. Except for gear graphs in the above list of star like graphs, there are precise formulas available in the literature to compute the inverse/Moore-Penrose inverse of their distance matrices. These formulas tell that if $D$ is the distance matrix of $G$, then $D^\dagger = -\frac{1}{2}L+uu'$, where $L$ is a Laplacian-like matrix which is positive semidefinite and all row sums equal to zero. The matrix $L$ and the vector $u$ depend only on the degree and number of vertices in $G$ and hence, can be written directly from $G$. The earliest formula obtained is for distance matrices of trees in Graham and Lovász \cite{GL}. In this paper, we obtain an elegant formula of this kind to compute the Moore-Penrose inverse of the distance matrix of a gear graph.

math.CO↗
arXiv2022

Learning Temporal Point Processes for Efficient Retrieval of Continuous Time Event Sequences

Recent developments in predictive modeling using marked temporal point processes (MTPP) have enabled an accurate characterization of several real-world applications involving continuous-time event sequences (CTESs). However, the retrieval problem of such sequences remains largely unaddressed in literature. To tackle this, we propose NEUROSEQRET which learns to retrieve and rank a relevant set of continuous-time event sequences for a given query sequence, from a large corpus of sequences. More specifically, NEUROSEQRET first applies a trainable unwarping function on the query sequence, which makes it comparable with corpus sequences, especially when a relevant query-corpus pair has individually different attributes. Next, it feeds the unwarped query sequence and the corpus sequence into MTPP guided neural relevance models. We develop two variants of the relevance model which offer a tradeoff between accuracy and efficiency. We also propose an optimization framework to learn binary sequence embeddings from the relevance scores, suitable for the locality-sensitive hashing leading to a significant speedup in returning top-K results for a given query sequence. Our experiments with several datasets show the significant accuracy boost of NEUROSEQRET beyond several baselines, as well as the efficacy of our hashing mechanism.

cs.IR↗
arXiv2022

Doing More with Less: Overcoming Data Scarcity for POI Recommendation via Cross-Region Transfer

Variability in social app usage across regions results in a high skew of the quantity and the quality of check-in data collected, which in turn is a challenge for effective location recommender systems. In this paper, we present Axolotl (Automated cross Location-network Transfer Learning), a novel method aimed at transferring location preference models learned in a data-rich region to significantly boost the quality of recommendations in a data-scarce region. Axolotl predominantly deploys two channels for information transfer, (1) a meta-learning based procedure learned using location recommendation as well as social predictions, and (2) a lightweight unsupervised cluster-based transfer across users and locations with similar preferences. Both of these work together synergistically to achieve improved accuracy of recommendations in data-scarce regions without any prerequisite of overlapping users and with minimal fine-tuning. We build Axolotl on top of a twin graph-attention neural network model used for capturing the user- and location-conditioned influences in a user-mobility graph for each region. We conduct extensive experiments on 12 user mobility datasets across the U.S., Japan, and Germany, using 3 as source regions and 9 of them (that have much sparsely recorded mobility data) as target regions. Empirically, we show that Axolotl achieves up to 18% better recommendation performance than the existing state-of-the-art methods across all metrics.

cs.LG↗
arXiv2021

Learning Neural Models for Continuous-Time Sequences

The large volumes of data generated by human activities such as online purchases, health records, spatial mobility etc. are stored as a sequence of events over a continuous time. Learning deep learning methods over such sequences is a non-trivial task as it involves modeling the ever-increasing event timestamps, inter-event time gaps, event types, and the influences between events -- within and across different sequences. This situation is further exacerbated by the constraints associated with data collection e.g. limited data, incomplete sequences, privacy restrictions etc. With the research direction described in this work, we aim to study the properties of continuous-time event sequences (CTES) and design robust yet scalable neural network-based models to overcome the aforementioned problems. In this work, we model the underlying generative distribution of events using marked temporal point processes (MTPP) to address a wide range of real-world problems. Moreover, we highlight the efficacy of the proposed approaches over the state-of-the-art baselines and later report the ongoing research problems.

cs.LG↗
arXiv2021

Region Invariant Normalizing Flows for Mobility Transfer

There exists a high variability in mobility data volumes across different regions, which deteriorates the performance of spatial recommender systems that rely on region-specific data. In this paper, we propose a novel transfer learning framework called REFORMD, for continuous-time location prediction for regions with sparse checkin data. Specifically, we model user-specific checkin-sequences in a region using a marked temporal point process (MTPP) with normalizing flows to learn the inter-checkin time and geo-distributions. Later, we transfer the model parameters of spatial and temporal flows trained on a data-rich origin region for the next check-in and time prediction in a target region with scarce checkin data. We capture the evolving region-specific checkin dynamics for MTPP and spatial-temporal flows by maximizing the joint likelihood of next checkin with three channels (1) checkin-category prediction, (2) checkin-time prediction, and (3) travel distance prediction. Extensive experiments on different user mobility datasets across the U.S. and Japan show that our model significantly outperforms state-of-the-art methods for modeling continuous-time sequences. Moreover, we also show that REFORMD can be easily adapted for product recommendations i.e., sequences without any spatial component.

cs.LG↗
arXiv2021

BERT Meets Relational DB: Contextual Representations of Relational Databases

In this paper, we address the problem of learning low dimension representation of entities on relational databases consisting of multiple tables. Embeddings help to capture semantics encoded in the database and can be used in a variety of settings like auto-completion of tables, fully-neural query processing of relational joins queries, seamlessly handling missing values, and more. Current work is restricted to working with just single table, or using pretrained embeddings over an external corpus making them unsuitable for use in real-world databases. In this work, we look into ways of using these attention-based model to learn embeddings for entities in the relational database. We are inspired by BERT style pretraining methods and are interested in observing how they can be extended for representation learning on structured databases. We evaluate our approach of the autocompletion of relational databases and achieve improvement over standard baselines.

cs.CL↗
arXiv2020

Modeling Implicit Communities using Spatio-Temporal Point Processes from Geo-tagged Event Traces

The location check-ins of users through various location-based services such as Foursquare, Twitter, and Facebook Places, etc., generate large traces of geo-tagged events. These event-traces often manifest in hidden (possibly overlapping) communities of users with similar interests. Inferring these implicit communities is crucial for forming user profiles for improvements in recommendation and prediction tasks. Given only time-stamped geo-tagged traces of users, can we find out these implicit communities, and characteristics of the underlying influence network? Can we use this network to improve the next location prediction task? In this paper, we focus on the problem of community detection as well as capturing the underlying diffusion process and propose a model COLAB based on Spatio-temporal point processes in continuous time but discrete space of locations that simultaneously models the implicit communities of users based on their check-in activities, without making use of their social network connections. COLAB captures the semantic features of the location, user-to-user influence along with spatial and temporal preferences of users. To learn the latent community of users and model parameters, we propose an algorithm based on stochastic variational inference. To the best of our knowledge, this is the first attempt at jointly modeling the diffusion process with activity-driven implicit communities. We demonstrate COLAB achieves up to 27% improvements in location prediction task over recent deep point-process based methods on geo-tagged event traces collected from Foursquare check-ins.

cs.SI↗