• Publications
  • Influence
Embedding Multimodal Relational Data for Knowledge Base Completion
TLDR
This paper proposes multimodal knowledge base embeddings (MKBE) that use different neural encoders for this variety of observed data, and combines them with existing relational models to learnembeddings of the entities and multi-modal data.
Investigating Robustness and Interpretability of Link Prediction via Adversarial Modifications
TLDR
An efficient approach to estimate the effect of adversarial modifications for link prediction models: identifying the fact to add into or remove from the knowledge graph that changes the prediction for a target fact after the model is retrained is introduced.
Revisiting Evaluation of Knowledge Base Completion Models
TLDR
A semi-complete KG is gathered, using a random subgraph from the test and validation data of YAGO3-10, which enables us to compute accurate triple classification accuracy on this data, and the shortcomings of these evaluation metrics are studied.
ParsiNLU: A Suite of Language Understanding Challenges for Persian
TLDR
This work introduces ParsiNLU, the first benchmark in Persian language that includes a range of language understanding tasks—reading comprehension, textual entailment, and so on, and presents the first results on state-of-the-art monolingual and multilingual pre-trained language models on this benchmark and compares them with human performance.
Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models
TLDR
Evaluation of OpenAI's GPT models, Google-internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters finds that model performance and calibration both improve with scale, but are poor in absolute terms.
An Empirical Comparison of Instance Attribution Methods for NLP
TLDR
It is found that simple retrieval methods yield training instances that differ from those identified via gradient-based methods (such as IFs), but that nonetheless exhibit desirable characteristics similar to more complex attribution methods.
Generating User-friendly Explanations for Loan Denials using GANs
TLDR
A novel Generative Adversarial Network (GAN) that can accommodate smaller datasets, to generate user-friendly textual explanations that help educate the loan applicants, or help them take appropriate action towards a future approval.
Combining Feature and Instance Attribution to Detect Artifacts
TLDR
This paper proposes new hybrid approaches that combine saliency maps (which highlight important input features) with instance attribution methods (which retrieve training samples influential to a given prediction) and shows that this proposed training-feature attribution can be used to efficiently uncover artifacts in training data when a challenging validation set is available.
Embedding Multimodal Relational Data
TLDR
This work proposes a new approach to represent relational triples, consisting of a subject entity, relation, and an object entity, by estimating fixed, low-dimensional representations for each entity and relation from observations, thus encode the uncertainty and infer missing facts accurately and efficiently.
Optimal tradeoff between source and state distortions over a Gaussian channel using single and hybrid digital analog codes
TLDR
The problem of transmitting an analog Gaussian source over an additive white Gaussian noise (AWGN) channel in the presence of a Gaussian interference known only at the transmitter is investigated and different transmission schemes based on joint source-channel coding are presented.
...
...