CASTELLO. The fifth session of the cycle of Smart Business Innovation & Values webinars, organized by the Castellón Chamber of Commerce, gave way yesterday afternoon to a fruitful dialogue on the status of the application of one of its branches -the 'Machine Learning' (ML)- in today's company, in a virtual meeting under the title The last stretch of Artificial Intelligence (AI). Both the reflections expressed in the first part, centered on a dialogue between the writer and scientific disseminator Adolfo Plasencia, and the CEO of BigML, Francisco J. Martín, pivoted around this axis, as well as in the subsequent round table -How can impact the AI in your company-, which included the participation of Vicent Botti, professor of Computing and director of the Valencian Research Institute for Artificial Intelligence -VRAIN- at the Polytechnic University of Valencia; Juan Ignacio Torregrosa, general director for Digital Advancement of the Generalitat Valenciana and Alfredo Cebrián, founding partner and CEO of Cuatroochenta. Sergio Méndez, director of the Panamerican Business School, was in charge of moderating the entire webinar.
With the premium sponsorship of CaixaBank, the session opened with a welcome by the director of the Chamber, Javier Valls, who thanked those attending the virtual meeting for their participation, as well as the collaboration of all the sponsors of the cycle and of the session .
Plasencia opened the initial dialogue with a presentation in which he explained that ML "is the branch of AI that is being implemented more quickly" and recalled that in essence, a software system "learns when its performance improves with experience and through the use of data”. Likewise, the writer highlighted that this technology “can be applied horizontally to a multitude of sectors and fields in any type of business, management and commercial operation, but what the entrepreneur wants to know is how he can use AI and ML in his company with results real”. Recalling his recent interview in Castellón Plaza, Plasencia emphasized that the AI vanguard "works with the third culture: the fusion of the techno-scientific culture with the humanistic culture, and this should also be taken into account by companies and regulators, because their social impact is enormous and ethical and social dilemmas arise in its application”.
Among other examples of the application of AI, the popularizer spoke of a recent investigation in which the Moncofin chemist Avelino Corma participates with researchers from MIT, in which ML has been applied to the synthesis of zeolites (minerals), reaching the conclusion that this method (machine learning) is well suited “to replace trial and error approaches traditionally used to guide experimentation”, calling ML “the missing link for high-performance discovery”. The but: the predictive models developed by this means “could not be interpretable by humans”, according to Corma, so “we have to continue working on the representation of data in a way that maximizes what humans and machines learn from each other. ”. Plasencia concluded that AI "is closer than we think, in companies."
In the exchange with Plasencia, Martín confirmed that the great advance in hardware in recent decades has not been accompanied by a significant progression in software, "and the first reference to Deep Learning in the scientific literature dates back to the mid-1980s" . The CEO of BigML lamented the “excessive optimism” registered in recent years around AI “just because certain things have worked well”. In this regard, it was clear: "I do not 'see' AI, as techniques that can become an application, for another 100 or 200 years: fundamental engineering is lacking for things to work." In this regard, and questioned by Botti, who had said that this difficulty "we already experienced when the software crisis", Martín agreed with him on the need to apply "engineering concepts to software to make it more productive". Thus, he concluded: “we have already experienced this and it is solvable: we have to see how more people can be trained… we need more software architects focused on AI solutions”.
Together with Plasencia, Martín reviewed his own career, which led him to create the technology companies Isoco and Strands -from which he sold more than 30 patents to Apple- before co-founding BigML in 2011. Over time, he detailed, "the Basic AI technique has evolved only marginally, so you have to cover it with a bunch of rules that have nothing to do with AI." According to Martín, "by combining ML with very basic reasoning we can have better results, and Europe should become stronger in this type of technique or combination of them". More specifically, in Spain he located "a very good opportunity for entrepreneurs who identify common needs of all companies in a specific sector, standardizing the data format that these companies consume, to provide solutions that provide a fundamental impact on these companies: for example , I am sure that many hotels and restaurants, in such a touristic country, do not have any short-term demand forecasting system, to manage their stock or the type of food that is going to be ordered based on the weather tomorrow or next week, and there is no software package that 50,000 Spanish restaurants can buy”. The businessman praised "the high level at which students are trained in Spain", which means that the only thing missing is "the spark to take risks at a business level, you just need to give them a little push".
In his case, BigML offers a more accessible, easier and cheaper ML tool for everyone, designed "to democratize it". According to Martin, it was about making the whole experience "enjoyable." Indian publication Analytics India Mag ranked the company's solution #1 among ML platforms internationally in 2021, highlighting its ability to build "a machine learning or deep learning model in just 3-4 clicks."
The challenges of AI The round table analyzed the current state of AI, the challenges facing society in the face of advances in this field and its future impact on human beings. Vicent Botti opened the fire with a reflection at the geostrategic level: “Europe has to make a significant effort not to miss the boat against China and the US, and a large part of the effort must be in the training of our AI professionals, to be more competitive” . For his part, Juan Ignacio Torregrosa predicted an “exponential” growth of its effects: “it is as if Edison had been asked about the effect of electricity; everyone had a suspicion that it was going to transform the world, and that's what can happen with AI."
How to Delete your Usage History Tracks in Windows https://t.co/JikJ9poEBl
— Kevin Huyn Tue May 24 00:42:52 +0000 2016
Linking his reflections with those presented by Martín in the first part of the session, Alfredo Cebrián highlighted that the change in recent years has been based "on the fact that the availability of data has accelerated and we have experienced a commoditization of the service" . The CEO of Cuatroochenta stated that "talking about AI is too broad an umbrella", so he preferred to talk about "simple models applicable in very specific contexts, which are effective in reducing costs in specific companies: a simple application of ML well applied, choosing its variables well, it modifies an entire sector”. Cebrián rejected the idea of embracing AI "in general", advocating that companies "do not lose their focus, ask themselves what value they create for their clients and delve into that if they really want to improve." And for this, "everything happens by investing time and money, and hiring good professionals, because the company that gets up in the morning thinking about how to improve, obsessed with the value it brings to the client, AI will enter through the direct door. or indirectly through specific technological solutions”.
For his part, Juan Ignacio Torregrosa stated that data “is the new oil, the energy source on which the economy is based” and predicted the next leap in automatic data generation, “for example, through the Internet of Things (IoT)”. In this sense, the general director for the Advancement of Digital pointed out that "this volume of data will give rise to hardware to process and store it", to conclude that the growth of AI will benefit Humanity "especially if we are able to correctly anonymize data and have a layer of cybersecurity that protects us from the bandits of the 21st century”. At this point, Cebrián drew attention to the danger posed by the possible introduction of changes "that as end users we judge as positive but that can cause global damage", appealing to have "a little more vision and responsibility in the global impact as a society”.
Botti recalled that technology is neutral and appealed for its use by the industrial sector, a sector that, as he explained, “is also interested in the fact that, together with ML, its systems also incorporate the knowledge accumulated by each company in its field. Likewise, the professor underlined the need for all AI to be explainable, since "this is the key for it to be reliable".
Regarding the possibility of AI turning against humans, Martín put forward a reassuring argument: “humans have an intelligence that makes us capable of adapting to different situations… I can pick up the mobile and the cup without needing to know their exact weight or which joint should I use: for the robots to adapt to that and react in real time, there is a long way to go. And that is not well transmitted in the media, there is everything to build. In this line, Torregrosa stressed: “our brain has no competition in the machines; the question is how much time would we save if many tasks that could be automated were to do so, that could improve our quality of life”.
Regarding the impact of digitization on employment, the participants at the table appealed to distinguish work (task) from job and salary, and also referred to the high demand for certain positions, both new profiles ("people trained in AI”, as Botti proposed) as well as other more traditional ones (“mathematicians”, Plasencia explained). At this point, Botti appealed to the importance of training society now "so that it understands what AI is, that it knows what it is" and in parallel, "create basic profiles of VT in AI, because in the Valencian Community it can have an immediate impact that type of training”, as well as postgraduate training (master's and doctorate). In this regard, the professor welcomed the launch by the Generalitat of the Valencian Graduate School Research Network Artificial Intelligence (ValgrAI), a postgraduate school and a network of research centers in Artificial Intelligence that aims to promote cooperation between the best teams in the Community and high-level training in AI.
Elaborating on the importance of education, Torregrosa indicated that its value is especially relevant in a productive fabric based on SMEs and micro-SMEs, and indicated that the Secretary of State for Digitization and AI "is going to carry out a very important program for this type of company , to advance in this field”. Likewise, the General Directorate headed by Torregrosa “has programs to financially support companies that present a digitization plan, through the IVF, and to train 100,000 citizens in AI, so that people lose their fear and consider this technology as their own. technology".
For his part, Cebrián advocated “letting the universities work, taking pressure off them, because a person will change their way of working several times in their working life: we have lost basics in education, what I want in my company is people who are dreamers, with ideas that make me change the way I think... today we are surrounded by information and we can spend an afternoon thinking, and that is a hyper-privileged position compared to 100 years ago. Sometimes we have a wrong defeatist feeling, because we have a great heritage.
The desirability of more disclosure about digitization floated throughout the session. Torregrosa stressed the need for new technologies not to exclude anyone due to the dangers that broad layers of society "disconnect" from the democratic system entail. And the tiptoe was Francisco J. Martín, asked by Plasencia about Spain's position in relation to digitization: “there are very interesting successes, and it's a shame that ordinary people don't know about them: there is too much motivation with football and politicians and very few minutes of TV dedicated to a technology that is going to impact the majority of people in the country, this will have to change one day.