Information bottleneck is an information-theoretic principle of representation learning that aims to learn a maximally compressed representation that preserves as much information about labels as possible. Under this principle, two different methods have been proposed, i.e., information bottleneck (IB) and deterministic information bottleneck (DIB), and have gained significant progress in explaining the representation mechanisms of deep learning algorithms. However, these theoretical and empirical successes are only valid with the assumption that training and test data are drawn from the same distribution, which is clearly not satisfied in many real-world applications. In this paper, we study their generalization abilities within a transfer learning scenario, where the target error could be decomposed into three components, i.e., source empirical error, source generalization gap (SG), and representation discrepancy (RD). Comparing IB and DIB on these terms, we prove that DIB's SG bound is tighter than IB's while DIB's RD is larger than IB's. Therefore, it is difficult to tell which one is better. To balance the trade-off between SG and the RD, we propose an elastic information bottleneck (EIB) to interpolate between the IB and DIB regularizers, which guarantees a Pareto frontier within the IB framework. Additionally, simulations and real data experiments show that EIB has the ability to achieve better domain adaptation results than IB and DIB, which validates the correctness of our theories.

Explaining the black-box predictions of NLP models naturally and accurately is an important open problem in natural language generation. These free-text explanations are expected to contain sufficient and carefully-selected evidence to form supportive arguments for predictions. Due to the superior generative capacity of large pretrained language models, recent work built on prompt engineering enables explanation generation without specific training. However, explanation generated through single-pass prompting often lacks sufficiency and conciseness. To address this problem, we develop an information bottleneck method EIB to produce refined explanations that are sufficient and concise. Our approach regenerates the free-text explanation by polishing the single-pass output from the pretrained language model but retaining the information that supports the contents being explained. Experiments on two out-of-domain tasks verify the effectiveness of EIB through automatic evaluation and thoroughly-conducted human evaluation.

Bias is a common problem inherent in recommender systems, which is entangled with users' preferences and poses a great challenge to unbiased learning. For debiasing tasks, the doubly robust (DR) method and its variants show superior performance due to the double robustness property, that is, DR is unbiased when either imputed errors or learned propensities are accurate. However, our theoretical analysis reveals that DR usually has a large variance. Meanwhile, DR would suffer unexpectedly large bias and poor generalization caused by inaccurate imputed errors and learned propensities, which usually occur in practice. In this paper, we propose a principled approach that can effectively reduce bias and variance simultaneously for existing DR approaches when the error imputation model is misspecified. In addition, we further propose a novel semi-parametric collaborative learning approach that decomposes imputed errors into parametric and nonparametric parts and updates them collaboratively, resulting in more accurate predictions. Both theoretical analysis and experiments demonstrate the superiority of the proposed methods compared with existing debiasing methods.

The European Investment Bank (EIB) is the long-term lending institution of the European Union owned by its Member States. It makes long-term finance available for sound investment in order to contribute towards EU policy goals. In 2019, the Bank provided over €850 million in loans for Danish projects in various sectors, including services, innovative industry and small and medium-sized enterprises. Under Horizon 2020, the EU research and innovation framework program for 2014-2020, the Infectious Diseases Finance Facility (IDFF) provides financial products ranging from standard debt to equity-type financing for amounts typically between €7.5 million and €75 million to innovative players active in developing innovative vaccines, drugs, medical and diagnostic devices or novel research infrastructure for combating infectious diseases. Project costs may include clinical trial costs, set-up of commercialization such as market access, development of prototypes or industrial roll-out of novel equipment, pre-clinical R&D costs and working capital requirement. This facility is delivered directly by the EIB, which has so far provided €241 million under the InnovFin IDFF. About Evaxion Evaxion Biotech A/S is an AI-immunology platform company decoding the human immune system to discover and develop novel, targeted immunotherapies to treat cancer and infectious diseases. Based on its proprietary and scalable AI-immunology core technology, Evaxion Biotech is developing a broad pipeline of novel assets which comprises three patient-specific cancer immunotherapies, two of which are in Phase 1/2 clinical development. In addition, the Company is advancing a portfolio of vaccines to prevent bacterial and viral infections with one program currently in preclinical development against S. aureus and skin and soft tissue infections (SSTI).

* EIB signs €20 million (DKK 149 million) loan agreement with Danish research and development company Evaxion Biotech A/S. * Evaxion will use financing for research and development into its proprietary artificial intelligence platforms used to identify novel vaccine and immunotherapy targets. * Transaction supported by European Commission through InnovFin – EU finance for innovators program, under the Infectious Diseases Finance Facility window. LUXEMBOURG and COPENHAGEN, Denmark, Dec. 09, 2020 (GLOBE NEWSWIRE) -- The European Investment Bank (EIB) has signed a €20 million loan agreement with Evaxion, supported by the Infectious Diseases Finance Facility window under the EU’s InnovFin – EU finance for innovators program, backed by the EU’s research and innovation program Horizon 2020, which guarantees EIB loans for highly innovative projects. The financing will support Evaxion’s research and development (R&D) investments in the use of artificial intelligence (AI) to determine the most suitable targets for immunotherapies for cancer and infectious diseases.The main focus of the research, Evaxion’s proprietary AI-based platforms, could potentially transform the drug discovery process for vaccines and immunotherapies. In particular, the infectious disease programs targeting Staphylococcus aureus and Pseudomonas aeruginosa aim to identify novel vaccines to address the resistance to common treatments developed by these pathogens. The EU financing will support the investments in Evaxion’s AI platforms, as well as its pre-clinical and clinical activity.Lars Wegner, CEO of Evaxion, said: “This loan agreement with the EIB underpins Evaxion’s position as a leader in the utilization of AI for discovery and development of novel vaccines and immunotherapies, and supports important research and development within our key programs. We are proud to be entering into this agreement with EIB, which is an example of critical financing being used to address some of the biggest challenges to human health.”EIB Vice-President Christian Kettel Thomsen stated: “As the EU bank, the EIB strives to ensure that high impact projects receive the financing they need. The EIB is delighted to be supporting Evaxion. The use of AI in the development of vaccines and immunotherapies is true innovation and is perfectly in sync with the EIB’s overarching priority: improving people’s lives. In recent years, the Bank’s investments in Denmark have been able to support many innovative companies, of which this is an excellent example.” Mariya Gabriel, European Commissioner for Innovation, Research, Culture, Education and Youth, added: “Protecting the health of our citizens and leading the digital transformation are key priorities underpinning Europe’s prosperity and growth. This EU-backed loan for Evaxion’s research using AI may help to transform the drug discovery process to treat cancer and infectious diseases.”The project not only strengthens European R&D, but is also contemplated to have an impact on the safeguarding and creation of highly skilled jobs in the EU. The research will be wholly carried out in Denmark, and is expected to contribute to technological advancements in artificial intelligence and its applications in healthcare. As such, the project is expected to bring important social benefits in terms of advances in research and alleviating the burden placed on healthcare services by the increased incidence of multi-resistant pathogens.Background Information:The European Investment Bank (EIB) is the long-term lending institution of the European Union owned by its Member States. It makes long-term finance available for sound investment in order to contribute towards EU policy goals. In 2019, the Bank provided over €850 million in loans for Danish projects in various sectors, including services, innovative industry and small and medium-sized enterprises.Under Horizon 2020, the EU research and innovation framework program for 2014-2020, the Infectious Diseases Finance Facility (IDFF) provides financial products ranging from standard debt to equity-type financing for amounts typically between €7.5 million and €75 million to innovative players active in developing innovative vaccines, drugs, medical and diagnostic devices or novel research infrastructure for combating infectious diseases. Project costs may include clinical trial costs, set-up of commercialization such as market access, development of prototypes or industrial roll-out of novel equipment, pre-clinical R&D costs and working capital requirement. This facility is delivered directly by the EIB, which has so far provided €241 million under the InnovFin IDFF.About Evaxion Evaxion Biotech A/S is an AI-immunology™ platform company decoding the human immune system to discover and develop novel, targeted immunotherapies to treat cancer and infectious diseases. Based on its proprietary and scalable AI-immunology core technology, Evaxion Biotech is developing a broad pipeline of novel assets which comprises three patient-specific cancer immunotherapies, two of which are in Phase 1/2 clinical development. In addition, the Company is advancing a portfolio of vaccines to prevent bacterial and viral infections with one program currently in preclinical development against S. aureus and skin and soft tissue infections (SSTI).Press contacts: European Investment Bank: Tim Smit, +352 691 286423, t.smit@eib.org European Commission: Flora Matthaes, +32 460 755148, flora.matthaes@ec.europa.eu Evaxion: Glenn S. Vraniak +1 (513) 476-2669, gvr@evaxion-biotech.com

The European Investment Bank today (November 29th) agreed a EUR 30 million facility to support investment by Dublin based biotech company, Nuritas, to further increase and accelerate development of artificial intelligence and DNA analysis to improve global healthcare. Representing the first direct EIB financing for an Irish biotech company, the new support will allow the company to access tranched financing as required to scale up the development of the rapidly growing company. Future investment by the company backed by the EIB will scale up development of new therapeutics in areas of interest including anti-aging, anti-inflammatory and diabetes treatment. "Nuritas has already demonstrated through PeptAIde, their first product launch this year, how the discovery of healthcare products can be rapidly accelerated using artificial intelligence. During due diligence over recent months the EIB has been impressed by their innovative use of technology and the number of global deals already achieved and they now become the first Irish biotech firm to benefit from the EIB's dedicated European Growth Finance Facility. Our goal is to find innovative and fast-growing companies so that we ensure they can benefit from access to long-term financing under the new initiative and Nuritas fits this bill perfectly. We look forward to working with other innovative and fast growing companies to ensure they can benefit from access to long-term financing under the new initiative. This new EUR 30 million support from the European Investment Bank will enable Nuritas to expand and commercialise even faster." said Andrew McDowell, European Investment Bank Vice President.

The company announced it has received €30 million ($36 million) funding from the European Investment Bank, the lending arm of the European Union. "We're proud to receive this stamp of approval from the EIB. It's the type of offer you can't refuse and it will allow us to further accelerate our growth and continue to level the playing field for small businesses, giving them access to tools to take on the big corporations," said Jacob de Geer, CEO and co-founder of iZettle, in a statement. The funding follows the startup's most recent round, which was earlier this year, when it raised $63 million at the same $500 million valuation it had in its last equity round. It appears the valuation is staying the same with this latest round: the money, as with earlier this year, is coming in the form of debt funding and will be distributed over the next three years, the company said.

Swedish payment terminal company iZettle has won €30 million (£26.6 million) in funding from the European Investment Bank (EIB) to explore artificial intelligence (AI) and machine learning for small businesses. AI and machine learning knowledge and technology are often geared towards big companies, and small businesses don't usually have the funds to develop solutions tailored to their needs. To tackle this problem, Stockholm-based iZettle plans to invest in research and development over the next three years, specifically to benefit smaller businesses. "We're proud to receive this stamp of approval from the EIB," said CEO and cofounder of iZettle Jacob de Geer. "It's the type of offer you can't refuse and it will allow us to further accelerate our growth and continue to level the playing field for small businesses, giving them access to tools to take on the big corporations," he said.