arXiv subjects

Ankit Chadha

Publications and source records attributed to Ankit Chadha.

17 recordsLinked to original sources

arXiv2026

Routed Graph Handoff: Adaptive Format Selection for Multi-Agent LLM Delegation

Multi-agent LLM systems coordinate through natural-language messages that consume 40--60\% of their token budget. Replacing these with structured graphs reduces cost but fails on tasks requiring adaptive reasoning. We propose \textbf{Routed Graph Handoff}, where a lightweight LLM router (155 tokens, 0.15\% overhead) selects between a typed dependency graph and natural language for each delegation. On four benchmarks (1,050+ trajectories), the routed system matches or exceeds NL-only on every task: \textbf{+12.7\,pp} on $\tau$-retail at 3.2$\times$ compression ($p{<}0.01$), \textbf{+8.7\,pp} on BrowseComp at 2.2$\times$ compression ($p{<}0.05$), and parity on BFCL and AppWorld. Without the router, graph-only delegation regresses 14.6\,pp on AppWorld; the router eliminates this at near-zero cost. A graph-aware executor prompt is required: the same schema without interpretation guidance yields no gain. An oracle analysis reveals 8.6\,pp of additional headroom, motivating execution-time adaptive routing as future work.

cs.CL↗

arXiv2026

APEX-MEM: Agentic Semi-Structured Memory with Temporal Reasoning for Long-Term Conversational AI

Large language models still struggle with reliable long-term conversational memory: simply enlarging context windows or applying naive retrieval often introduces noise and destabilizes responses. We present APEX-MEM, a conversational memory system that combines three key innovations: (1) a property graph which uses domain-agnostic ontology to structure conversations as temporally grounded events in an entity-centric framework, (2) append-only storage that preserves the full temporal evolution of information, and (3) a multi-tool retrieval agent that understands and resolves conflicting or evolving information at query time, producing a compact and contextually relevant memory summary. This retrieval-time resolution preserves the full interaction history while suppressing irrelevant details. APEX-MEM achieves 88.88% accuracy on LOCOMO's Question Answering task and 86.2% on LongMemEval, outperforming state-of-the-art session-aware approaches and demonstrating that structured property graphs enable more temporally coherent long-term conversational reasoning.

cs.CL↗

arXiv2026

APEX-EM: Non-Parametric Online Learning for Autonomous Agents via Structured Procedural-Episodic Experience Replay

LLM agents rerun full reasoning for every task, even one they solved moments earlier. We introduce \textbf{APEX-EM}, a non-parametric experience memory that stores complete procedural-episodic traces in a typed Procedural Knowledge Graph (PKG) and retrieves them through three channels: semantic search, structural-signature matching over abstract operation sequences, and graph traversal. A Plan-Retrieve-Generate-Iterate-Ingest (PRGII) workflow produces, quality-gates, and commits experiences, indexing both successes and failures so the agent learns what to reuse and what to avoid. No weights change during deployment. We evaluate on five benchmarks: BigCodeBench, KGQAGen-10k, HLE, Lifelong Agent Bench, and ALFWorld. Because prior work uses different backbones, we base our claims on same-backbone comparisons that hold model capability fixed. On held-out BigCodeBench transfer with a shared GPT-4o backbone, APEX-EM gains +7.6\,pp over the no-memory baseline, $3.3\times$ MemRL's +2.3\,pp under the identical setup. On Lifelong Agent Bench with a shared GPT-4o-mini backbone, it gains +1.4\,pp (OS) and +1.0\,pp (DB) cumulative success. On KGQAGen-10k, frozen memory transfers to a blind 1{,}079-question test split at 73.7\% versus 42.0\% with no memory, approaching an oracle handed the ground-truth subgraph (84.9\%). Across three Opus scales the memory gain stays at +27 to +32\,pp, so it adds to model capability rather than substituting for it. Component analysis shows no single mechanism dominates: teacher feedback is negligible for code but adds +10.3\,pp on structured queries, structural signatures give $3.3\times$ the transfer of semantic-only retrieval, and within-epoch iteration recovers most of the gain when rich feedback is unavailable. These results argue for modular memory composed per domain.

cs.CL↗

arXiv2023

Question-Context Alignment and Answer-Context Dependencies for Effective Answer Sentence Selection

Answer sentence selection (AS2) in open-domain question answering finds answer for a question by ranking candidate sentences extracted from web documents. Recent work exploits answer context, i.e., sentences around a candidate, by incorporating them as additional input string to the Transformer models to improve the correctness scoring. In this paper, we propose to improve the candidate scoring by explicitly incorporating the dependencies between question-context and answer-context into the final representation of a candidate. Specifically, we use Optimal Transport to compute the question-based dependencies among sentences in the passage where the answer is extracted from. We then represent these dependencies as edges in a graph and use Graph Convolutional Network to derive the representation of a candidate, a node in the graph. Our proposed model achieves significant improvements on popular AS2 benchmarks, i.e., WikiQA and WDRASS, obtaining new state-of-the-art on all benchmarks.

cs.CL↗

arXiv2023

Controlled Text Generation with Hidden Representation Transformations

We propose CHRT (Control Hidden Representation Transformation) - a controlled language generation framework that steers large language models to generate text pertaining to certain attributes (such as toxicity). CHRT gains attribute control by modifying the hidden representation of the base model through learned transformations. We employ a contrastive-learning framework to learn these transformations that can be combined to gain multi-attribute control. The effectiveness of CHRT is experimentally shown by comparing it with seven baselines over three attributes. CHRT outperforms all the baselines in the task of detoxification, positive sentiment steering, and text simplification while minimizing the loss in linguistic qualities. Further, our approach has the lowest inference latency of only 0.01 seconds more than the base model, making it the most suitable for high-performance production environments. We open-source our code and release two novel datasets to further propel controlled language generation research.

cs.CL↗

arXiv2023

Cross-Lingual Knowledge Distillation for Answer Sentence Selection in Low-Resource Languages

While impressive performance has been achieved on the task of Answer Sentence Selection (AS2) for English, the same does not hold for languages that lack large labeled datasets. In this work, we propose Cross-Lingual Knowledge Distillation (CLKD) from a strong English AS2 teacher as a method to train AS2 models for low-resource languages in the tasks without the need of labeled data for the target language. To evaluate our method, we introduce 1) Xtr-WikiQA, a translation-based WikiQA dataset for 9 additional languages, and 2) TyDi-AS2, a multilingual AS2 dataset with over 70K questions spanning 8 typologically diverse languages. We conduct extensive experiments on Xtr-WikiQA and TyDi-AS2 with multiple teachers, diverse monolingual and multilingual pretrained language models (PLMs) as students, and both monolingual and multilingual training. The results demonstrate that CLKD either outperforms or rivals even supervised fine-tuning with the same amount of labeled data and a combination of machine translation and the teacher model. Our method can potentially enable stronger AS2 models for low-resource languages, while TyDi-AS2 can serve as the largest multilingual AS2 dataset for further studies in the research community.

cs.CL↗

arXiv2022

ACM -- Attribute Conditioning for Abstractive Multi Document Summarization

Abstractive multi document summarization has evolved as a task through the basic sequence to sequence approaches to transformer and graph based techniques. Each of these approaches has primarily focused on the issues of multi document information synthesis and attention based approaches to extract salient information. A challenge that arises with multi document summarization which is not prevalent in single document summarization is the need to effectively summarize multiple documents that might have conflicting polarity, sentiment or subjective information about a given topic. In this paper we propose ACM, attribute conditioned multi document summarization,a model that incorporates attribute conditioning modules in order to decouple conflicting information by conditioning for a certain attribute in the output summary. This approach shows strong gains in ROUGE score over baseline multi document summarization approaches and shows gains in fluency, informativeness and reduction in repetitiveness as shown through a human annotation analysis study.

cs.CL↗

arXiv2022

Training Mixed-Domain Translation Models via Federated Learning

Training mixed-domain translation models is a complex task that demands tailored architectures and costly data preparation techniques. In this work, we leverage federated learning (FL) in order to tackle the problem. Our investigation demonstrates that with slight modifications in the training process, neural machine translation (NMT) engines can be easily adapted when an FL-based aggregation is applied to fuse different domains. Experimental results also show that engines built via FL are able to perform on par with state-of-the-art baselines that rely on centralized training techniques. We evaluate our hypothesis in the presence of five datasets with different sizes, from different domains, to translate from German into English and discuss how FL and NMT can mutually benefit from each other. In addition to providing benchmarking results on the union of FL and NMT, we also propose a novel technique to dynamically control the communication bandwidth by selecting impactful parameters during FL updates. This is a significant achievement considering the large size of NMT engines that need to be exchanged between FL parties.

cs.CL↗

arXiv2021

Communication-Efficient Federated Learning for Neural Machine Translation

Training neural machine translation (NMT) models in federated learning (FL) settings could be inefficient both computationally and communication-wise, due to the large size of translation engines as well as the multiple rounds of updates required to train clients and a central server. In this paper, we explore how to efficiently build NMT models in an FL setup by proposing a novel solution. In order to reduce the communication overhead, out of all neural layers we only exchange what we term "Controller" layers. Controllers are a small number of additional neural components connected to our pre-trained architectures. These new components are placed in between original layers. They act as liaisons to communicate with the central server and learn minimal information that is sufficient enough to update clients. We evaluated the performance of our models on five datasets from different domains to translate from German into English. We noted that the models equipped with Controllers preform on par with those trained in a central and non-FL setting. In addition, we observed a substantial reduction in the communication traffic of the FL pipeline, which is a direct consequence of using Controllers. Based on our experiments, Controller-based models are ~6 times less expensive than their other peers. This reduction is significantly important when we consider the number of parameters in large models and it becomes even more critical when such parameters need to be exchanged for multiple rounds in FL settings.

cs.CL↗

arXiv2019

Deep Reinforced Self-Attention Masks for Abstractive Summarization (DR.SAS)

We present a novel architectural scheme to tackle the abstractive summarization problem based on the CNN/DMdataset which fuses Reinforcement Learning (RL) withUniLM, which is a pre-trained Deep Learning Model, to solve various natural language tasks. We have tested the limits of learning fine-grained attention in Transformers to improve the summarization quality. UniLM applies attention to the entire token space in a global fashion. We propose DR.SAS which applies the Actor-Critic (AC) algorithm to learn a dynamic self-attention distribution over the tokens to reduce redundancy and generate factual and coherent summaries to improve the quality of summarization. After performing hyperparameter tuning, we achievedbetter ROUGE results compared to the baseline. Our model tends to be more extractive/factual yet coherent in detail because of optimization over ROUGE rewards. We present detailed error analysis with examples of the strengths and limitations of our model. Our codebase will be publicly available on our GitHub.

cs.CL↗

arXiv2019

BERTQA -- Attention on Steroids

In this work, we extend the Bidirectional Encoder Representations from Transformers (BERT) with an emphasis on directed coattention to obtain an improved F1 performance on the SQUAD2.0 dataset. The Transformer architecture on which BERT is based places hierarchical global attention on the concatenation of the context and query. Our additions to the BERT architecture augment this attention with a more focused context to query (C2Q) and query to context (Q2C) attention via a set of modified Transformer encoder units. In addition, we explore adding convolution-based feature extraction within the coattention architecture to add localized information to self-attention. We found that coattention significantly improves the no answer F1 by 4 points in the base and 1 point in the large architecture. After adding skip connections the no answer F1 improved further without causing an additional loss in has answer F1. The addition of localized feature extraction added to attention produced an overall dev F1 of 77.03 in the base architecture. We applied our findings to the large BERT model which contains twice as many layers and further used our own augmented version of the SQUAD 2.0 dataset created by back translation, which we have named SQUAD 2.Q. Finally, we performed hyperparameter tuning and ensembled our best models for a final F1/EM of 82.317/79.442 (Attention on Steroids, PCE Test Leaderboard).

cs.CL↗

arXiv2015

Dual-Layer Video Encryption using RSA Algorithm

This paper proposes a video encryption algorithm using RSA and Pseudo Noise (PN) sequence, aimed at applications requiring sensitive video information transfers. The system is primarily designed to work with files encoded using the Audio Video Interleaved (AVI) codec, although it can be easily ported for use with Moving Picture Experts Group (MPEG) encoded files. The audio and video components of the source separately undergo two layers of encryption to ensure a reasonable level of security. Encryption of the video component involves applying the RSA algorithm followed by the PN-based encryption. Similarly, the audio component is first encrypted using PN and further subjected to encryption using the Discrete Cosine Transform. Combining these techniques, an efficient system, invulnerable to security breaches and attacks with favorable values of parameters such as encryption/decryption speed, encryption/decryption ratio and visual degradation; has been put forth. For applications requiring encryption of sensitive data wherein stringent security requirements are of prime concern, the system is found to yield negligible similarities in visual perception between the original and the encrypted video sequence. For applications wherein visual similarity is not of major concern, we limit the encryption task to a single level of encryption which is accomplished by using RSA, thereby quickening the encryption process. Although some similarity between the original and encrypted video is observed in this case, it is not enough to comprehend the happenings in the video.

cs.CR↗

arXiv2014

ARC Sort: Enhanced and Time Efficient Sorting Algorithm

This paper discusses about a sorting algorithm which uses the concept of buckets where each bucket represents a certain number of digits. A two dimensional data structure is used where one dimension represents buckets i. e; number of digits and each bucket's corresponding dimensions represents the input numbers that belong to that bucket. Each bucket is then individually sorted. Since every preceding bucket elements will always be smaller than the succeeding buckets no comparison between them is required. By doing this we can significantly reduced the time complexity of any sorting algorithm used to sort the given set of inputs.

cs.DS↗

arXiv2013

Biometric Signature Processing & Recognition Using Radial Basis Function Network

Automatic recognition of signature is a challenging problem which has received much attention during recent years due to its many applications in different fields. Signature has been used for long time for verification and authentication purpose. Earlier methods were manual but nowadays they are getting digitized. This paper provides an efficient method to signature recognition using Radial Basis Function Network. The network is trained with sample images in database. Feature extraction is performed before using them for training. For testing purpose, an image is made to undergo rotation-translation-scaling correction and then given to network. The network successfully identifies the original image and gives correct output for stored database images also. The method provides recognition rate of approximately 80% for 200 samples.

cs.CV↗

arXiv2013

Image Steganography using Karhunen-Loeve Transform and Least Bit Substitution

As communication channels are increasing in number, reliability of faithful communication is reducing. Hacking and tempering of data are two major issues for which security should be provided by channel. This raises the importance of steganography. In this paper, a novel method to encode the message information inside a carrier image has been described. It uses Karhunen-Loève Transform for compression of data and Least Bit Substitution for data encryption. Compression removes redundancy and thus also provides encoding to a level. It is taken further by means of Least Bit Substitution. The algorithm used for this purpose uses pixel matrix which serves as a best tool to work on. Three different sets of images were used with three different numbers of bits to be substituted by message information. The experimental results show that algorithm is time efficient and provides high data capacity. Further, it can decrypt the original data effectively. Parameters such as carrier error and message error were calculated for each set and were compared for performance analysis.

cs.MM↗

arXiv2013

Orthogonal Frequency Division Multiplexing and its Applications

Orthogonal Frequency Division Multiplexing (OFDM) is a multi-carrier modulation technique which is very much popular in new wireless networks of IEEE standard, digital television, audio broadcasting and 4G mobile communications. The main benefit of OFDM over single-carrier schemes is its ability to cope with severe channel conditions without complex equalization filters. It has improved the quality of long-distance communication by eliminating InterSymbol Interference (ISI) and improving Signal-to-Noise ratio (SNR). The main drawbacks of OFDM are its high peak to average power ratio and its sensitivity to phase noise and frequency offset. This paper gives an overview of OFDM, its applications in various systems such as IEEE 802.11a, Digital Audio Broadcasting (DAB) and Digital Broadcast Services to Handheld Devices (DVB-H) along with its advantages and disadvantages.

cs.NI↗

arXiv2013

An Efficient Method for Image and Audio Steganography using Least Significant Bit (LSB) Substitution

In order to improve the data hiding in all types of multimedia data formats such as image and audio and to make hidden message imperceptible, a novel method for steganography is introduced in this paper. It is based on Least Significant Bit (LSB) manipulation and inclusion of redundant noise as secret key in the message. This method is applied to data hiding in images. For data hiding in audio, Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) both are used. All the results displayed prove to be time-efficient and effective. Also the algorithm is tested for various numbers of bits. For those values of bits, Mean Square Error (MSE) and Peak-Signal-to-Noise-Ratio (PSNR) are calculated and plotted. Experimental results show that the stego-image is visually indistinguishable from the original cover-image when n<=4, because of better PSNR which is achieved by this technique. The final results obtained after steganography process does not reveal presence of any hidden message, thus qualifying the criteria of imperceptible message.

cs.MM↗